"Asure is committed to offering our clients the best tools and resources to support their growth," said Asure Chairman and CEO, Pat Goepel. "By integrating Lendio into our marketplace, we're ensuring that our clients have direct access to top-tier financing solutions that can propel their businesses forward."

Asure Software and Lendio Announce Strategic Partnership to Launch Lendio in the Asure Marketplace

As health systems continue to deal with workforce shortages, many leaders are turning to technology to try to alleviate some of the burdens on their existing staff. The Guthrie Clinic, based in Sayre, Pa., created a virtual care command center, dubbed the ...

How health systems are using tech to address staff shortages

For additional information on these measures and reconciliations to the most directly comparable GAAP measures, see “Non-GAAP Financial Measures” and Table 2 to the “Financial Tables” herein. Technology Platform segment total net revenue includes ...

SoFi Technologies, Inc. Reports Third Quarter 2023 Results

For additional information on these measures and reconciliations to the most directly comparable GAAP measures, see "Non-GAAP Financial Measures" and Table 2 to the "Financial Tables" herein. Technology Platform segment total net revenue includes ...

Although English is not his first language, Whitehouse High School senior Sebastian Toro doesn’t feel excluded from classes or interacting with his peers. And it’s all due to

Whitehouse ISD benefits from technology that provides real-time translation for students

With its innovative use of technology, the bank has transformed the way it delivers banking services to its customers ... and provides customers with real-time access to their account information. This has not only made banking more convenient for ...

The Role of Technology in Transforming Banking Services

Coming together in this way will help us to further establish our position as the market leader in the foodservices’ industry” said Group President Aharon Shamash of Informa Connect North America. “With strong brands and an unmatched combined database of unique contacts,

Informa Connect Announces Formation of Food Services Group

Fintech, short for financial technology, represents the marriage of finance and technology to improve financial services. It encompasses a wide ... As financial transactions and personal information become increasingly digital, the potential for data ...

The Rise of Fintech: How Banking Software Technology is Reshaping the Financial Landscape

Connecting Level's technology with Cox's fiber-powered network and 25 years of experience in the multifamily space ensures industry-leading connectivity and speed and gives property owners and renters a fully integrated service experience they won't find ...

Cox Selects Level as Next-Generation IoT Partner and Begins Deployments to Cox Customers

Delivered strong performance with 3Q23 financial results ~~ Increasing full year 2023 outlook ~PLANO, Texas, Oct. 26, 2023 (GLOBE NEWSWIRE)

Integer Holdings Corporation Reports Third Quarter 2023 Results

Table XI

tablexi

128 Stories To Learn About Charles Dickens

This paper is under CC BY 4.0 DEED license. available on arxiv Authors: (1) Ehsan Toreini, University of Surrey, UK; (2) Maryam Mehrnezhad, Royal Holloway University of London; (3) Aad Van Moorsel, Birmingham University. Table of Links Abstract & Introduction Background and Related Work FaaS Architecture Implementation and Performance Analysis Conclusion Acknowledgment and References 3 FaaS Architecture In this Section, we present the architecture of our system (Fig. 1) and describe its features. The FaaS architecture includes stakeholders in three roles: A) ML System: a system that owns the data and the ML algorithm, B) Fairness Auditor Service: a service that computes the fair performance of the ML system, and C) Universal Verifier: anyone who has the technical expertise and motivation to verify the auditing process. 3.1 Threat Model The design and implementation of the security of parties implementing the respective protocol roles (ML system, Fairness Auditor Service, and Universal Verifier) (Fig. 1) are independent of each other. The inter–communications that happen between the roles assumes no trust between parties; thus, all their claims must be accompanied with validation proofs (for which we will use ZKP). We assume the Auditor System is vulnerable to different attacks and not trustworthy. Thus, the data stored on the Fairness Auditor System must be encrypted, tamper-proof and verifiable at all stages. Moreover, we assume the communication channel between the ML system and fairness auditor is not protected. Therefore, the sensitive data must be encrypted before the transmission starts. However, there will be an agreement on the cryptographic primitives at the pre–setting stage in the protocol sequence. In FaaS, we assume that the ML system is honest in sending the cryptograms of the original labels of the dataset samples. One might argue against such assumption and discuss that the ML system might intend to deceive the Auditor Service, and by extension the verifiers, by modifying the actual labels of the dataset. For instance, the ML system would provide the cryptograms of the actual labels and the predicted ones as similar to each other as possible so that the auditor concludes the algorithms are fair. This is an interesting area for further research. For instance, it may be addressed by providing the cryptograms of the actual labels to the Auditor Service independently e.g. the verifier may own a dataset it provides to a ML system. The verifier then separately decides the desired values for the actual labels and feeds these to the Auditor service. In this way, it is far less clear to the ML system how to manipulate the data it sends to the auditor, since some of the labels come from elsewhere. The internal security of the roles is beyond FaaS. The ML system itself needs to consider extra measures to protect its data and algorithms. We assume the ML system does present the data and predictions honestly. This is a reasonable assumption since the incentives to perform ethically is in contrast to being dishonest when participating in fairness auditing process. This is discussed more in the Discussion Section. Table 2: Possible permutations of 3-bit representation of an entry in the original data. 3.2 Protocol Overview The main security protocol sequence is between the ML system and Fairness Auditing Service or auditor in short form. Note that although we suggest three roles in our architecture, the communications are mainly between the above two roles, and any universal verifier can turn to the auditor service (which represents the fairness board), if they want to challenge the computations. The ML system is responsible for the implementation and execution of the ML algorithm. It has data as input and performs some prediction (depending on the use case and purpose) that forms the output (Fig. 1). The Fairness Auditor Service receives information from the ML system, evaluates its fairness performance by computing a fairness metric. Then, it returns the result for the metric back to the ML system. It also publishes the calculations in a fairness board for public verification. The public fairness board is a publicly accessible, read-only fairness board (e.g. a website). The auditor only has the right to append data (and the sufficient proofs) to the fairness board. Also, the auditor verifies the authenticity, correctness and integrity of data before publishing it. 3.3 Protocol Sequence This protocol has three stages: setup, cryptogram generation and fairness metric computation. 3.3.1 Phase I: Setup In this phase, the ML System and Auditor agree on the initial settings. We assume the protocol functions in multiplicative cyclic group setting (i.e. Digital Signature Algorithm (DSA)–like group [18]), but it can also function in additive cyclic groups (i.e. Elliptic Curve Digital Signature Algorithm (ECDSA)–like groups [18]). The auditor and ML system publicly agree on (p, q, g) before the start of the protocol. Let p and q be two large primes where q|(p − 1). In a multiplicative cyclic group (Z ∗ p ), Gq is a subgroup of prime order q and g is its generator. For simplicity, we assume the Decision Diffie–Hellman (DDH) problem is out of scope [31]. Next, the ML system generates a public/private pair key by using DSA or ECDSA and publishes the public keys in the fairness board. The protection the private key pair depends on the security architecture of the ML system and we assume the private key is securely stored in an industrial standard practice (e.g. using the secure memory module on board). After initial agreements, the ML system produces a cryptogram table with n rows corresponding to the number of samples in their test dataset. We will refer to this table as cryptogram table in the rest of this paper. In case the ML system does not want to reveal the number of the samples in the test set, the auditor and the ML system can publicly agree on n. In this case, n must be big enough so that the universal verifiers are satisfied with the outcome. Cryptogram Table: Each row in the cryptogram table summarises three parameters: (1) protected group membership status, (2) its actual label and (3) predicted label by the ML model. Each row contains the encrypted format of the three parameters along with proofs of its correctness. A cryptogram table in the setup phase is shown in Table 3. In the simplest case, each parameter is binary. Therefore, the combined parameters will generate eight permutations in total. In the setup phase, the table is generated to contain all eight possible permutations and their proofs for each data sample. The total structure of the permutations are shown in Table 2. Each row will satisfy four properties: (a) one can easily verify if a single cryptogram is the encrypted version of one of the eight possible permutations, (b) while verifiable, if only one single cryptogram selected, one cannot exert which permutations the current cryptogram represents, (c) for each two cryptograms selected from a single row, anyone will be able to distinguish each from one another, and (d) given a set of cryptograms arbitrarily select from each row as a set, one can easily check how many cases for each “permutation” are in the set. The generation of the cryptogram table functions are based on the following sequence: Step (1): For each of the n samples, the system generates a random public key g xi where xi is the private key and xi ∈ [1, q − 1]. Step (3): The corresponding column number that equals the decimal value of the binary encoding is selected from the cryptogram table to complete the fairness auditing table( as shown in Table 2). Finally, the generated fairness auditing table is digitally signed by the ML system and then is sent over the Fairness auditing service. 3.3.3 Phase III: Fairness Evaluation First, the fairness auditing service receives the fairness auditing table, verifies the digital signature and the ZKPs, and publishes the contents in the fairness board. At this point, we expand each of these equation components to compare them together. This process is computationally heavy especially when the number of data samples in the fairness auditing table is large. In this case, the fairness auditor can delegate the declaration of the permutation number to the ML system. The auditor still receives the fairness auditing table and the relevant ZKPs. It can store the fairness auditing table to the fairness board, compute the fairness, and verify the correctness of the declared permutation numbers. The universal verifier can follow the same steps to verify the fairness metric computations through the fairness auditing table that is publicly accessible via fairness board. At the end of this stage, the auditor uses the acquired numbers to compute the fairness metric and release the information publicly. The number of each permutation denotes the overall performance of the ML algorithm for each of the groups with protected attribute. Table 4 demonstrates the permutations and how it relates to the fairness metric of the ML system. The cryptogram table and the results will be published on the fairness board (Fig. 1)

FaaS Architecture and Verifiable Fairness for ML Systems

FaaS আর্কিটেকচার এবং ML সিস্টেমের জন্য যাচাইযোগ্য ন্যায্যতা

FaaS-Architektur und nachweisbare Fairness für ML-Systeme

Arquitectura FaaS y equidad verificable para sistemas de aprendizaje automático

Architecture FaaS et équité vérifiable pour les systèmes ML

एमएल सिस्टम के लिए FaaS आर्किटेक्चर और सत्यापन योग्य निष्पक्षता

FaaS アーキテクチャと ML システムの検証可能な公平性

ML 시스템을 위한 FaaS 아키텍처 및 검증 가능한 공정성

Arquitetura FaaS e justiça verificável para sistemas de ML

Архитектура FaaS и проверяемая честность для систем машинного обучения

ML Sistemleri için FaaS Mimarisi ve Doğrulanabilir Adalet

Kiến trúc FaaS và tính công bằng có thể kiểm chứng cho hệ thống ML

ML 系统的 FaaS 架构和可验证公平性

Top Software Development Companies 2022 That Are Growing Fast

Get to know the top software outsourcing company and 29 more runners-up in this latest market research by Explority. Last year, the global outsourcing market at 85.6 billion US dollars. A notable part of this market, outsourcing software development is not subsiding either. was valued at The reasons behind this upward trend are understandable: according to the , businesses turn to outsourcing as it brings cost optimization and makes them more agile, efficient, and effective. Deloitte Global Outsourcing Survey 2018 With the growing popularity of outsourcing as a service, hundreds of IT companies started positioning it as one of their core capabilities and strengths. But do all of these firms prove to be so skilled in delivering outsourcing software development? To help you define who’s who in today’s outsourcing market, we have assembled the list of the most trustworthy vendors ready to handle software development, UI/UX design, and disruptive technologies. To make the list objective, we ranked the companies according to the following criteria: technological strength ability to ensure project security focus on innovation diversity and complexity of completed projects geographical flexibility Below, we present 30 outsourcing software development companies that have showed the highest scores in these criteria over the last 12 months. 1. Iflexion Founded in 1999, Iflexion houses 850+ experts in outsourcing software development, integration, business analysis, QA, and project management. Iflexion delivers such services as custom software development for end business users, product development, and DevOps enablement, among others. The company also provides dedicated development teams and is ready to boost their software solutions with emerging technologies like AI, AR, and big data, should the project require it. Iflexion has worked with such market giants as Philips, Expedia, Toyota, and adidas, with more than 1,500 completed projects for clients in 30+ countries overall. Location: Denver, CO Founded: 1999 Major Clients: Google, eBay, PayPal, Philips, Expedia, Toyota, adidas Number of Employees: 50-249 Industry Presence: IT, Hospitality, Retail, Media, Finance Website: https://www.iflexion.com/ 2. 10Pearls With more than 350 mobile developers, software developers, designers, and researchers worldwide, 10Pearls augments businesses’ capabilities with their technological expertise. 10Pearls creates applications driven by such emerging technologies as artificial intelligence, blockchain, and augmented reality. Location: Vienna, VA Founded: 2004 Major Clients: National Geographic, Johnson & Johnson, AARP, Nextel Number of Employees: 250-999 Industry Presence: Healthcare, Media, Finance, Transportation Website: https://10pearls.com/ 3. Orases Since 2000, Orases has been delivering strategy, branding, interface design, and full-stack outsourcing software development to more than 750 clients from ten countries all around the world. The company expertizes in building products to be integrated with other systems: ERP, CRM, EMR, and others. Orases also delivers branding services, including brand, communication and social media strategies. Location: Frederick, MD Founded: 2000 Major Clients: NFL, Johns Hopkins University, Frederick County Bank, Major League Baseball Number of Employees: 10-49 Industry Presence: Healthcare, Finance, Sports Website: https://orases.com/ 4. Trigent Software US-based Trigent Software has been providing outsourcing software development  since 1995. The company delivers consulting, design, and development across cloud, mobility, business intelligence, product engineering, and emerging tech domains. Over the years, the company have designed more than 400 products adopted in large and small enterprises alike. Location: Fayville, MA Founded: 1995 Major Clients: Navistar, NEPC, Honeywell, Hewlett-Packard, American Heart Association Number of Employees: 250-999 Industry Presence: Finance, Healthcare, IT, Education, Transportation Website: https://www.trigent.com/ 5. Spiria Spiria is a Canadian custom software development company that has been delivering outsourcing software development for more than 15 years. Spiria offers web, mobile and UX/UI design development. Spiria specializes in user-friendly and responsive websites, from ecommerce stores to secure intranets. The company also delivers embedded systems, IoT solutions, and M2M solutions. Location: Montreal, Canada Founded: 2003 Major Clients: Autodesk, Shutterstock, Disney, Paymi, DogWatch, Evenko Number of Employees: 50-249 Industry Presence: IT, Education, Manufacturing Website: https://www.spiria.com/ 6. Fingent With an active client base across four continents and 250+ specialists on board, Fingent remains a competitive outsourcing software development company for over 15 years. They offer enterprise software development, web application development, data analytics, and digital transformation consulting. Location: White Plains, NY Founded: 2003 Major Clients: Sony, Johnson & Johnson, University of North Carolina, PwC, CBN Number of Employees: 50-249 Industry Presence: Transportation, Real Estate, Finance Website: https://www.fingent.com/ 7. STX Next Poland-based STX Next has 170+ Python, React Native, JavaScript, Android, iOS, and full-stack developers ready to take up outsourcing software development. The company also offers product design and software testing. STX Next offers end-to-end development, team extension, and project takeover. It also works with startups, helping them enter the market and scale throughout product lifecycles. Location: Poznań, Poland Founded: 2005 Major Clients: Seerene, Verve, Pixomondo, Unity, Lamden Number of Employees: 250-999 Industry Presence: Finance, IT, Real Estate Website: https://stxnext.com/ 8. Praxent Operating for more than 20 years, Praxent has delivered more than 300 software solutions in the areas of mobile and web applications development, web portals, and legacy software modernization. Besides, Praxent offers human-centered design development, branding, and rapid prototyping. Location: Austin, TX Founded: 2000 Major Clients: Cisco, Kinder Morgan, Sysco, World Market, Flex, Keller Williams Number of Employees: 50-249 Industry Presence: Finance, Healthcare, IT Website: https://praxent.com/ 9. The Software House With 129 specialists on board and 150+ successfully delivered projects, The Software House offers the following services in the niche of outsourcing software development: web and mobile development, digital product design, QA, and DevOps services. The Software House’s technological expertise includes React, Node.js, Angular, iOS, Android, and React Native. Location: Gliwice, Poland Founded: 2012 Major Clients: Smartum, Synerise, eSky Number of Employees: 50-249 Industry Presence: Finance, Manufacturing, Healthcare Website: https://tsh.io/ 10. AgileEngine Over the past nine years, AgileEngine has delivered more than 70 projects in outsourcing software development, including risk modeling and case management software, tracking platforms, and BI tools. Besides UI, web, mobile, and enterprise software development, AgileEngine provides QA services that include load, stress, performance, security, and usability testing. Location: McLean, VA Founded: 2010 Major Clients: Needles, Logi Analytics, Farmers Edge, LivingSocial Number of Employees: 250-999 Industry Presence: IT, Healthcare Website: https://agileengine.com/ 11. Highland Solutions Established in 1999, Highland Solutions is a Chicago-based custom software development company that engineers and integrates web and mobile applications, multi-tenant custom platforms, and CRM systems. Highland Solutions also offers development of OutSystems low-code platforms that automate time-consuming and monotonous manual tasks. Location: Chicago, IL Founded: 1999 Major Clients: Thrivent Financial, The Delta Institute, Morton Arboretum, Creative Group, Harvard University Health Services Number of Employees: 10-49 Industry Presence: Finance, Manufacturing, Education, Healthcare, IT Website: https://highlandsolutions.com/ 12. Merixstudio Since 1999, Merixstudio has provided outsourcing software development to more than 200 customers from a variety of market segments. The software development company specializes in web development, Python programming, frontend development, React Native mobile development, and UX/UI design. Location: Poznań, Poland Founded: 1999 Major Clients: Fox, The Independent, DFJ, Divio Number of Employees: 50-249 Industry Presence: Finance, Education, Real Estate, Sport Website: https://www.merixstudio.com/ 13. Caxy Interactive Serving outsourcing software development from its Chicago-based office, Caxy Interactive expertizes in Drupal, JavaScript, and Python development, ecommerce solutions, and machine learning. The software development company also offers strategy development and Agile consulting services. Besides, Caxy Interactive takes up UX and web design projects. Location: Chicago, IL Founded: 1999 Major Clients: International Code Council, The Field Museum, Rush University Medical Center, Uptake Technologies Number of Employees: 10-49 Industry Presence: Finance, Education, Manufacturing, IT Website: https://www.caxy.com/ 14. Boldare Founded in 2004, Boldare was formed after a merger of two software development companies, XSolve and Chilid. So far, the company’s team of 130+ technological experts have delivered 250+ projects. Boldare provides full-cycle outsourcing software development, prototyping, and design. It also offers digital transformation services. Location: Gliwice, Poland Founded: 2004 Major Clients: BlaBlaCar, Sonnen, Takamol, Holiday Taxis, Agnito Number of Employees: 50-249 Industry Presence: Energy, Media, Ecommerce, Healthcare Website: https://www.boldare.com/ 15. MojoTech US-located MojoTech outsourcing software development and engineering, UX/UI design, and digital transformation services. Besides, the company mentors their clients’ in-house IT teams. For more than 11 years of their market presence, MojoTech’s team of 65 delivered 150+ software development projects. Location: Providence, RI Founded: 2008 Major Clients: AccuWeather, Aetna, Groove, Blue Cross Blue Shield, United Way Number of Employees: 10-49 Industry Presence: Finance, Consumer Services, IT Website: https://www.mojotech.com/ 16. ClearSummit ClearSummit is a software development company offering mobile and web app development, frontend and backend development, and UI/UX design. The company also offers discovery and technical advisory consultations. Additionally, ClearSummit offers software maintenance and support services to help clients fix bugs and upgrade their systems. Location: Marina del Rey, CA Founded: 2014 Major Clients: Belkin, TuneRegistry, SCH Travel, Musea Number of Employees: 10-49 Industry Presence: Consumer Services, IT, Media, Real Estate Website: http://clearsumm.it 17. Future Processing Established in 2000, Future Processing have delivered 500 solutions to 150+ clients from all over the world. The company’s expertise in outsourcing software development includes software development and consulting using .NET, Java, React, Angular, and other programming languages. The software development company brings to its clients such emerging technologies as AI, blockchain, and big data analytics, and delivers software audit and business intelligence setup. Location: Gliwice, Poland Founded: 2000 Major Clients: Tenneco, Ista, Valeo, TechSoup, Thomas Miller Number of Employees: 250-999 Industry Presence: Finance, Healthcare, Transportation, Real Estate Website: https://www.future-processing.com/ 18. Moove It Being a leading outsourcing software development company, Moove It offers product discovery, web and mobile apps development, testing automation, as well as digital transformation services. Besides these, the company provides technological and Agile coaching. Location: Austin, TX Founded: 2006 Major Clients: Unilever, Verifone, Disney, Sierra Wireless Number of Employees: 50-249 Industry Presence: Consumer Services, Education, Finance, Healthcare, IT Website: https://moove-it.com/ 19. iTexico Since its inception in 2010, iTexico has been delivering UI/UX design, software engineering, QA, mobile and cloud development from its six US-based offices. To make their outsourcing software development as efficient as possible, iTexico follows Agile, DevOps, and Lean methodologies. Location: Austin, TX Founded: 2010 Major Clients: Undisclosed Number of Employees: 50-249 Industry Presence: IT, Healthcare, Manufacturing Website: https://www.itexico.com/ 20. Django Stars European company Django Stars has been delivering its software solutions for almost 11 years. The company offers web and mobile app development, design, and QA. Django Stars also expertizes in data science, data architecture, intelligent chatbots, and big data analytics. The company’s core strength is in its Python development expertise: Django Stars promises to create any type of product using this programming language. Location: Kiev, Ukraine Founded: 2008 Major Clients: MoneyPark, Sindeo, Molo Finance, Clear Minds Number of Employees: 50-249 Industry Presence: Finance, Transportation, Healthcare, Real Estate, Hospitality Website: https://djangostars.com 21. Spire Digital US-based Spire Digital boasts 300+ products launched and 600+ clients served to an overall 98% satisfaction rate. The company delivers digital transformation, Agile process optimization, CX research, data science consulting, and lean product development. Besides all of these and expertise in AI, blockchain, and IoT, Spire Digital offers digital marketing strategy development. Location: Denver, CO Founded: 1998 Major Clients: TomTom, Lockheed Martin, US Navy, Helix, Western Union Number of Employees: 50-249 Industry Presence: Education, Finance, Healthcare, Ecommerce Website: https://www.spiredigital.com/ 22. KitelyTech Delivering to clients of all sizes, including startups and Fortune 500 companies, KitelyTech offers web and mobile app design, ecommerce solutions, data analytics consulting, QA, and more. The company offers SEO services as well. Among the emerging technologies that KitelyTech is offering, there are AI, blockchain, and AR. Location: Chicago, IL Founded: 2009 Major Clients: Karrot Health, Plastic Surgery Clinic of Chicago, Olympus, J.P. McMahon Number of Employees: 50-249 Industry Presence: Healthcare, IT, Transportation Website: https://kitelytech.com/ 23. Table XI US-based Table XI has been a trusted supplier of outsourcing software development for over 16 years. The company provides digital strategy and product design, along with mobile and custom web development. So far, the company have joined forces with at least 100 global clients. Location: Chicago, IL Founded: 2002 Major Clients: Dept. of Medicine of the Northwestern University, Journal for the American Medical Association, AccuWeather, Dickson Number of Employees: 10-49 Industry Presence: Healthcare, Manufacturing, Education Website: https://www.tablexi.com/ 24. Syberry Having a 94% client retention rate, Syberry serves at least 20 industries in the US alone. The company provides a whole range of services, including custom software development, QA, business analysis, and technology consulting. Syberry’s expertise also includes cloud, ERP, SaaS, big data, blockchain, and IoT. Location: Austin, TX Founded: 2014 Major Clients: Bombardier, PayPal, Disney, Shell, Livestream Number of Employees: 50-249 Industry Presence: Finance, Healthcare, Retail, Education, Energy, Transportation Website: https://www.syberry.com/ 25. Vincit California Vincit California is a 400+ employee company engineering mobile apps, web portals, and embedded systems. Besides, Vincit offers UX/UI design. The company has seven offices around the world, which makes it specifically fit for global software development outsourcing. Location: Irvine, CA Founded: 2007 Major Clients: Nokia, Logitech, Yamaha, Roadtrip Nation, Sigora, Fastems Number of Employees: 250-999 Industry Presence: IT, Finance, Telecommunications Website: https://www.vincit.com/ 26. Napoleon IT UK-located DCSL offers outsourcing software development, software and hardware integration, IT consulting, and ongoing system support. The company’s other services include team augmentation, data migration, legacy software modernization, and others. Location: Reston, VA Founded: 2011 Major Clients: Unilever, Severstal, Russian Railways, Yamaha Number of Employees: 50-249 Industry Presence: Retail, Ecommerce, Healthcare, Consumer Services Website: https://napoleonit.com/ 27. DCSL Software UK-located DCSL offers software development, software and hardware integration, IT consulting, and ongoing system support. The company’s other services include team augmentation, data migration, legacy software modernization, and others. Location: London, United Kingdom Founded: 1994 Major Clients: WaterAid, Mitsubishi, Hospify, Abbott, Lidl, Pearson Number of Employees: 50-249 Industry Presence: Healthcare, Education, Manufacturing, Real Estate, IT Website: https://www.dcslsoftware.com/ 28. DockYard Since 2010, DockYard has been delivering product strategy, UX, digital design, software development, software support, and maintenance services. DockYard also provides QA and testing, team augmentation, and business analysis. Location: Boston, MA Founded: 2010 Major Clients: Netflix, Apple, NASDAQ, Zipcar, Fidelity Investments, Harvard University Number of Employees: 50-249 Industry Presence: IT, Transportation, Education Website: https://dockyard.com 29. SoftwareMill SoftwareMill helps its clients tackle IT challenges with outsourcing software development, data science, system integration, and analytics. The company delivers enterprise solutions, UI/UX design, mobile and web applications, and DevOps integration. Location: Warsaw, Poland Founded: 2009 Major Clients: Tipser, Knip, Flexys Number of Employees: 50-249 Industry Presence: Finance, IT, Energy, Ecommerce Website: https://softwaremill.com/ 30. Tivix Specializing in Agile development of cloud-connected web and mobile applications, Tivix is a US-based company housing 65+ experienced developers across six offices worldwide. The company’s expertise in outsourcing software development includes API, database, Django, Angular, and React development. Location: San Francisco, CA Founded: 2008 Major Clients: Apple, Tesla, SolarCity, Zoetis, VMware Number of Employees: 50-249 Industry Presence: Finance, IT, Consumer Services, Healthcare What Does the Future Hold for Outsourcing Software Development Companies? As time passes and businesses start looking for new technologies such as cognitive computing and RPA, some of these vendors can start lagging or cease delivering outsourcing at all. To keep the listing current and help you keep tabs on in this ever-changing market, we’ll continue monitoring the best outsourcing software companies. Stay tuned. ) (Find out more about Explority and our ranking methodology, click here and here

Top 30 Outsourcing Software Development Companies in 2020

Someday you’ll come along with an idea to create the next big app which will bring your name into repute or make you rich. First of all, you can start googling “who can I hire to build an app”. Because Google knows everyone and everything. Or, if you have a similar experience, you’ll search “outsourcing software development company”. But of course, choosing might take a few weeks to a few months before picking that one for all your development requirements. Here is the list of trusted IT Outsourcing Companies based on Quora and Appfutura research. VironIT VironIT is a custom software development company with a primary focus on delivering end-to-end software development services. Company’s core areas of expertise are development of mobile applications, web-oriented software products, business software solutions, software integration, updating, support and maintenance of developed applications. VironIT has successfully carried out over 500 successful projects to over 300 high-profile companies in eHealth, Sports, Banking, Retail, Marketing, Gaming and other industries. The company’s clients range from SMEs to Fortune 500 corporations based in Western Europe and North America. VironIT’s team consists of 100+ seasoned specialists: software developers, engineers, and scientists with expertise in different technical domains. All of the company’s members have bachelor’s or master’s degrees in sciences and rich experience in the software development field. Most of VironIT’s employees are senior or mid-level developers. Ciklum Ciklum is a software solutions company headquartered in Kyiv, Ukraine, with offices in the UK, Denmark, Spain, Belarus, and the U.S., and more. With about 3,000 employees, Ciklum not only specializes in software engineering but also provides e-commerce and mobile app development, application testing, IT strategy consulting, and staff augmentation. Their clients range from large multinational companies to fast-growing startups working toward an IPO. Ciklum was hired by an enterprise IT firm as the main development resource: backend and frontend developers, quality assurance specialists, and business analysts. Operating as part of the team on an ongoing basis, Ciklum developed an e-commerce platform for the IT firm’s client that resulted in $50,000,000 per year in sales. Eight Bit Studios Founded in 2008, Eight Bit Studios is based in Chicago, IL. They focus on mobile apps, customer software development, and UX/UI design. The team of 35+ employees helps key industries, including education and gaming burst business. Eight Bit Studios redeveloped a product for an augmented-reality gaming platform. They had to review the platform to learn how as operating and then piece the platform back together. They were able to ensure all code worked properly. PSL Corp. PSL Corp. is a software engineering company that was founded in 1986. With locations in New York; Medellín, Colombia; and Mexico City, PSL Corp. has more than 650 employees. They specialize in custom software development, big data consulting, cloud consulting, system integration, and web development. PSL Corp. supports mid-sized and enterprise clients across a range of industries. PSL Corp. built a marketing web application for a sales software firm. The client valued PSL Corp.’s ability to program flawless, bug-free code thanks to their high level of expertise in development. To support the client and maintain quality assurance, PSL Corp. continues to provide ongoing maintenance and support. Table XI Table XI is a mobile app and custom software developer based in Chicago, IL. Founded in 2004, they are 35 dedicated app and website builders with a flair for bespoke design and digital strategy. Table XI recreated the website of a global events company, building a video player and an iOS app to deliver content, all with no site downtime. After the migration, the app and site functioned seamlessly. The client recommended Table XI for their cost, clarity, and adherence to deadlines. Intellectsoft Founded in 2007, Intellectsoft is a full-service software partner, focusing on app development for web, mobile, and emerging technologies. Headquartered in Palo Alto, California, their 200+ business strategists and tech specialists collaborate across seven global offices to create digital solutions for startups and enterprises alike. Intellectsoft created an e-commerce app with matching backend infrastructure, then integrated it into the client’s blockchain technology. Transforming the given creative demands into a satisfactory product under a strenuous deadline was deemed impressive, as were their communication skills and determination. iTechArt iTechArt is a top-tier, one-stop custom software development company. 73% of its clients come from referrals. With over 800 experienced engineers, iTechArt works with startups and other fast-growing companies to create the coolest, most innovative, and successful mobile and web applications. It also provides custom software development and QA & testing services in the industry. Itransition Founded in 1998 outside Denver, CO, Itransition now has an office in Austin, TX, as well as the UK and Belarus. They bring their 20 years experience and 1500+ person team to the table for custom software development, web development, and e-commerce development. Itransition was hired by another development firm to collaborate on a project for a tourism company. They migrated 44 websites into one e-commerce platform, Magento. 10Pearls Based in Vienna, VA, 10Pearls designs and maintains apps for various companies. The company and its nearly 350 employees have experience in cybersecurity, product development, and UX/UI design. A telemedicine company needed assistance with platform development and maintenance. 10Pearls helped the telemedicine company with all aspects of platform development, creating a website and mobile app and ensuring the website’s backend was admin-friendly. LOVATA LOVATA is an IT company headquartered in Minsk, Belarus. Founded in 2008, they provide services for web development and custom software development. They have about 24 employees that serve a range of clients in the healthcare, e-commerce, retail, and financial services sectors. LOVATA designed a website to increase a publishing company’s brand awareness. Revenue rose 400% as a result of more paid subscriptions and increased audience size. RubyGarage Development agency RubyGarage is located in Dnipro, Ukraine. Since it was founded in 2011, their team has evolved to about 100 employees. RubyGarages partners primarily with small businesses. Their services are centered on custom software, mobile app, and web development. RubyGarage provides digital marketing and development services for a software company. The client noted that the team’s development services were high quality. The marketing campaigns they executed improved traction and visibility during the entire length of the engagement. Belatrix Software Belatrix Software is an app development company headquartered in Argentina. The large team also has offices in Latin America, Florida, New York, and California. Founded in 2003, the agency is known as one of the top providers for software engineering services. Belatrix Software offered front-end development services to a financial group, helping the client’s existing development team build productivity-related tools. The client lacked the necessary development resources to process a backlog of work, so Belatrix Software’s development expertise complimented their in-house development team. The project is ongoing, and the client praised Belatrix Software’s work. Wizeline Headquartered in San Francisco, Wizeline is a product development company with offices in Guadalajara and Mexico City, Mexico, and Ho Chi Minh, Vietnam. Since its founding in 2014, the team has grown to over 390 employees. Wizeline specializes in custom software, web, and mobile app development and UX/UI design. An entertainment company hired Wizeline to improve their website’s CMS. Based on the team’s initial success, they fostered a partnership that’s still ongoing. The agency’s skills made them stand out from comparable firms throughout the process. The client appreciates Wizeline’s collaborative spirit and excellent communication skills. Fingent Technologies Founded in 2003, Fingent is a web and mobile development company which has offices in the United States, India and the United Arab Emirates. Its global team of 230 employees has helped hundreds of mid and large size organizations implement software solutions that increase productivity and profits. Oxagile Oxagile provides end-to-end custom software development services to SMBs and enterprises, focusing on online video management, advertising, real-time communication, and data solutions. The company’s headquarters are located in New York, with an EMEA office in London and the development center in Belarus. Oxagile offers both Scrum and Waterfall methodologies. With Vodafone among their clients and Microsoft among the partners, Oxagile is one of the top custom software development companies for video solutions for SMBs and enterprises. SumatoSoft SumatoSoft offers full-cycle app development, starting from the zero, through consulting, developing and finishing custom software. They work for large enterprises, startup companies, create B2B and B2C e-commerce, automated chatbot and more. SumatoSoft’s collaboration with Toyota was one of the biggest projects of the company. Here, they introduced new and optimized existing processes of the business, supported their marketing system, ensured an effective functioning of their offices across the world, rearranged and recreated the entire Toyota’s IT system. Praxent Praxent is not a simple custom software IT company. They have the role of an IT assistant to businesses of all sizes. By introducing custom cloud solutions to a ticket broker TicketCity, Praxent liberated 98% of their inventory, resulting in expansion of the system capacity by the whole 6800%. Moreover, Praxent gained 12,000 subscribers for Plumb’s veterinary drugs within a short time period, making the veterinary handbook easily accessible online for veterinarians across the world. Orases Orases is a modern agency from the US with standard services on their menu, including mobile and web applications development, custom software services and flexible IT solutions. They list SMBs and Fortune 500 companies as only some of the businesses they work with it. This company was involved in two industries that not many custom software IT outsource companies cover — sports and farming. Namely, they created websites for Dairy Farmers of America, Homegrown Frederick and Prince George’s County, as well as MLB Pitch Hit & Run, and several apps for NFL. FrogSlayer FrogSlayer provides growing companies with the custom software tools they need to accelerate growth and become the technology or data leader in their industry. Since 2011, its team has helped over 60 clients become technology leaders in a variety of industries. It has made commercial buildings smarter, oil rigs safer, financial crime harder, education more accessible, healthcare more efficient, and much more. FrogSlayer’s clients have generated over $350M in ROI — an average of 23x payback in 2 years or less. N-iX N-iX is an Eastern European IT company with over 16 years of excellence as a reliable technology partner for innovative companies across North America, UK, and Europe. N-iX has over 800 experienced software engineers working within development centers in Lviv, Kyiv (Ukraine), Kraków (Poland), Minsk (Belarus) and representative offices in Malmö (Sweden) and St. Petersburg, FL (USA). Hope this list of notable outsourcing companies will help you to burst your business to the next level. If you have seen any company that should be added to this list, then leave a note in comments.

Top IT Outsourcing Companies

“ , also known as paging, is the process of dividing a document into discrete pages, either electronic pages or printed pages.” Pagination It is most common technique to manage large data set at server/client side to distribute in chunks called as pages. In today’s time, Social media client apps improvised this by inventing “Infinite scroll”. Infinite scrolling allows users to load content continuously, eliminating the need for user’s explicit actions. App loads some initial data and then load the rest of the data when the user reaches the bottom of the visible content. This data is divided in pages. A framework for Infinite scrolling? As apps size and content, are increasing day by day, we need to implement infinite scroll in more than one places in our app. Although implementing infinite scroll is not very tricky but it is a repetitive job and create redundant code inside TableView/CollectionView and yes it breaks the “DRY” principal too. Whenever we have repeated code, there is always a chance of introducing a bug. I lately implemented infinite scroll in some of apps and I felt that at times you need to implement this in quite a few places, typically if you are working on a large e-commerce app. I was looking for a solution, a framework which does this job for me. I was not able to find a non-intrusive and easy to integrate solution where I don’t need to sub-class my ViewController or use a special kind of tableView/collectionView for this and It can work with my network stack. In short, I don’t want to create direct dependency of a framework/library all over in my project code. So I thought of giving it a try. Initial, I was not clear on and part. My refactored efforts were scattered all over the place. While I ended up having a non-intrusive solution (what I played out as a requirement), I learned an important lesson after making mistakes. what how Zoom out as far as you can until you are clear who is playing what role. Initially I was relentlessly trying to decouple/refactored code without identifying my interface and participant objects or classes. You can only do “ ” when you identify class is going to do . Separation of Concern which what What it takes to use your framework/Library When I started refactoring my code for library, I had not laid out my interfaces well. Generally this is a common problem. We usually don’t developed solutions which are being designed for larger audience but used within few use cases. It really open our eyes when we start thinking in this direction. And yes, this should be second step when we develop a framework. What part: 1. We need an Inter-actor/ Controller which Maintains number of visible rows Load next page at the end of scroll if required Maintain current page count and update all of these when it gets next page data array from the network. 2. To accomplish all of this, It needs some supporting function on top of tableView/collectionView 3. An interface between network stack. How Part: The trickiest part was the interfacing with network. It took some time to figure out how to separate it from the dependency of network stack. Most of pagination responses have the common structure which comprises of total Page count, current page and an array of elements. With this assumption, A can be defined for all responses over generic parameter Type. struct { types: [ ] page: totalPageCount: (types: [ ], page: , totalPageCount: ) { .types = types .page = page .totalPageCount = totalPageCount
    }
} public < > struct PageInfo T var T var Int var Int public init T Int Int self self self Now with this PageInfo struct in place, We have a generic way of giving data back to Inter-actor/Controller for updating the results after getting it from web service. This solves one problem of puzzle, another piece of puzzle is to indicate Network stack to download next pages data as and when required. So I designed an interface between Pageable and network stack which has two methods. { -> } public : protocol PageableService class /// Indicates when to load next page func load (pageInteractor: PageInteractor, page: Int) Void /// Cancel request if required func cancelAllRequests () Function is called by when it figure out that there is more data to load. A typical implementation can be LoadPage PageInteractor -> { resource: < <[ ]>> = ? prepareResource(page: page,
                                                                               pageSize: pageSize,
                                                                               pathForREST: ) { } info: < >?
    networkManager.fetch(res: resource) { (res) res { .success(result):
            info = (types: result.types,
                            page: result.page,
                            totalPageCount: result.totalPageCount) .failure(err): (err)
        } pgInteractor.returnedResponse(info)
    }
} func load (pageInteractor: PageInteractor, page: Int) Void guard let Resourse PagedResponse User try "/api/users" else return // Construct PageInfo to be returned var PageInfo User in switch case let PageInfo case let print // Require to call method on PageInteractor for updating result. First method indicates to request page. As It loads next set of data, the loaded data needs to be sent to PageInteractor for updating result. func loadPage( _ page: Int) -> Void when which Now calling , is a requirement which can be missed as API is not enforcing it. PageInteractor Completion blocks are here! Completion block helps in situations like this. So we can update an interface with completion block. As operates on generic parameter, This requires to create a generic interface to adopt by. PageInfo<T> { -> ) } public : protocol PageableService class <Item: Decodable> func loadPage ( page: Int,
                                   completion: @escaping _ (PageInfo<Item>?) Void func cancelAllRequests () This create a clear requirement for user to send data back to Page Interactor as well as gives a clean way to separate library from your networking code. Pageable Power of KeyPath: gives a feature where duplicate entries from server can be avoided at client side if somehow new entries are being added dynamically at server end. More can be found . Pageable here To use this feature your pagination data should have a unique Id in response array so that duplicates can be identified. To make this feature usable, It requires a user to give a power to provide a generic way to identify that field. Earlier due to inability to provide this information to , User needs to implement interface methods which checks items in dictionary and then add in final array to avoid duplication. This is very redundant task for user and ideally handled by only. PageInteractor PageDataSource PageInteractor { -> < > { startIndex = pgInteractor. () items = items ? [ ] { new items { pgInteractor.dict[ (new.id)] == {
                    pgInteractor.dict[ (new.id)] = (new.id)
                    pgInteractor.array.append(new) }
            }
        } startIndex..<pgInteractor. ()
    } { items = items ? [ ] {
            pgInteractor.array = items new items {
                pgInteractor.dict[ (new.id)] = (new.id)
            }
        }
    }
} : extension UserView PageDataSource func addUniqueItems ( items: [AnyObject]) for Range Int let count if let as User for in if String nil String String // adding items in array to be displayed return count func addAll (items: [AnyObject]) if let as User for in String String KeyPath are a way to reference properties without invoking them. It can be composed to make different path and used to get/set underlying properties. With the KeyPath passed in the initialiser of , implementation of both methods can move entirely in . This will relive user from not implementing . PageInteractor PageInteractor PageDataSource { -> < > { startIndex = () keypath = keypath { new items { key = new[keyPath: keypath] dict[key] == {
                    dict[key] = key
                    array.append(new)
                }
            }
        } startIndex..< ()
    } {
        array = items keypath = keypath { } new items { key = new[keyPath: keypath]
            dict[key] = key
        }
    }
    
} extension PageInteractor /**
     Server can add/remove items dynamically so it might be a case that
     an item which appears in previous request can come again due to
     certain element below got removed. This could result as duplicate items
     appearing in the list. To mitigate it, we would be creating a parallel dictionary
     which can be checked for duplicate items
     
     - Parameter items: items to be added
     - Parameter keypath: In case if duplicate entries has to be filter out,
     It requires keypath of unique items in model data.
     */ open func addUniqueFrom (items: [Element], keypath: KeyPath<Element, KeyType>?) Range Int let count if let for in let if nil return count /** Add all items, If there is empty list in table view
     - Parameter items: items to be added
     - Parameter keypath: In case if duplicate entries has to be filter out,
     It requires keypath of unique items in model data.
     */ open func addAll (items: [Element], keypath: KeyPath<Element, KeyType>?) guard let else return for in let With default implementation embedded in , All it takes is to provide KeyPath of unique data type to to use this feature. PageInteractor PageInteractor pageInteractor: < , > = (firstPage: firstReqIndex, service: service, keyPath: \UserModel.id) let PageInteractor UserModel Int PageInteractor Conclusion While developing infinite scrolling as a framework, was fun and turned out to be an easy solution to integrate with few steps to follow. I learned an important lesson on part. There is always a room of improvement when you look back to your solution. I did a few iterations while finalising interface for Pageable and now I feel next time I can avoid few of those. what You can find library at Github, feel free to give it a try/feedback ! here

TableView, CollectionView Infinite Scrolling: An Easy Way in Swift

Recently, I haven’t been updating my open-source articles as frequently — not because I’ve stopped writing, but because the progress on open-source commercialization has been great, and the endless task of drafting proposals has consumed my time. As a native open-source commercial company, WhaleOps employs mostly engineers. Asking these open-source contributors to write proposals wastes their development time, and their proposals don’t always meet the quality standard. Unlike managing in a big company, being a startup CEO means stepping into any role the company needs. After every strategic meeting, I’m the first to roll up my sleeves and tackle the most urgent tasks. As a result, I haven’t had time to write articles, as my limited time is mostly taken up with creating proposals that follow the formal template style. Especially recently, with one bid after another, I’ve found myself questioning my sanity. The existing large models can’t solve my problem, so I’ll build my own As a tech person, I always look to tools to solve my problems. Naturally, I thought of large models, but they can’t fully grasp the nuances of our products and often produce unreliable outputs. Plus, you’re the one delivering the work, not the model. So, I decided to develop a tool for proposal generation based on large models using Python and a compatible large model (compatible with ChatGPT). This tool automatically generates a proposal from your product documentation, breaking down product manuals into feature points. Based on a human-created mapping of these points to the requirements, it generates a Word version of the proposal and an Excel deviation table. The model can condense or expand content or simply copy relevant functionality as needed. Features of the open-source proposal tool The functionality of this tool is simple, with the most challenging part being the Word formatting (formatting in Word is always a pain). I experimented with several methods to make sure it follows the correct Title 1, Title 2, Title 3, body text, table, and image formats in the Word template. Staying true to the open-source spirit, I’ve uploaded the tool to my personal GitHub under an Apache License, so feel free to use it if you need. Here’s what it does: Get it here: https://github.com/William-GuoWei/ProposalLLM Breaks down your product manual into a set of reusable detail documents, reducing the need to repeatedly reference the source document when drafting a proposal. You can also customize specific functionality (the default file name is “Template.docx”).
Based on the requirements table filled in by a person, it automatically generates a proposal in a point-to-point response format, including all headings and content, with the correct Title 1, 2, 3 formatting, and automatically organizes body text, images, and bullet points (the default requirements table is “requirements_table.xlsx,” and the generated content is in “proposal_content.docx”).
For any product requirements in the corresponding functionality section, it automatically copies the product manual content into the point-to-point response section, retaining images, tables, and bullet points. You can also rewrite the product description to suit different proposal needs. If there’s no matching functionality, the model automatically generates relevant content (review and modify as needed).
Completes the technical requirements deviation table by automatically filling in responses in the “requirements_table.xlsx,” with responses formatted as “Answer: Fully supports, {Model-generated text based on project requirements}” and includes the section number corresponding to the proposal. With this tool, you can quickly modify and generate proposals at will. Proposal generation process Start by running Extract_Word.py to generate your product feature points as a Word document, then run Generate.py. If a feature point is missing, you can enter “X” in the Excel file, and the model will generate content that meets the requirements. However, I strongly recommend a manual review. After generation, you’ll see the proposal formatted with a table of contents, body text, images, tables, and bullet points, all automatically organized. The entire proposal is structured according to the client’s requirements in sequential format, with subheadings, content, images, and point-to-point responses. This takes care of all the repetitive work. The automatically generated deviation table includes everything, whether the content was generated by the model or not, along with the corresponding section numbers. You only need to finalize the deviation table with “&” symbols based on the final requirements — the model has written it all for you. For me, this tool has reduced what used to be 8 hours of work to around 30 minutes, and even our business team can generate the business proposal sections using the template. In total, this has cut down our time for a week-long proposal to 1–2 days, reducing the manpower required by 80%. How to use it? First, download all the code to a single directory from GitHub at https://github.com/William-GuoWei/ProposalLLM: Proposal Large Model (Chinese version). Install the Python environment and packages: pip install openpyxl, docx, openai, requests, docx, python-docx.
Apply for a ChatGPT or Baidu Qianfan large model key (I used ERNIE-Speed-8K, which is free), record the token, and place it in the relevant code section.
Copy your product manual to Template.docx. Be sure to use the body text, Title 1, Title 2, and Title 3 styles provided; other formats may cause issues.
Run Extract_Word.py to generate the feature point document from the product manual (supports up to 3 heading levels). If the list formatting appears off, don’t worry; the final format will align properly.
Fill in Columns B and C (which will generate secondary and tertiary headings automatically) and Column G (the corresponding product manual chapter). If a chapter is missing, enter “X.” Note that if there is no corresponding chapter or an “X” is entered, the model will generate content automatically.
Review the “proposal_content.docx” document and keep the chapter for which you want to start generating the proposal. You can modify the body text and heading 1, 2, and 3 styles; just don’t rename the styles, or there may be errors.
Adjust the parameters in Generate.py: API_KEY and SECRET_KEY: Baidu Cloud large model keys.
MAX_WIDTH_CM: Maximum image width; images larger than this will be resized automatically.
The prompts for generating content have been customized for large data scenarios, but feel free to modify them.
MoreSection=1 will read Column C to generate detailed tertiary headings (default is on).
ReGenerateText=0 will re-generate text content automatically for different proposal needs (default is off).
DDDAnswer=1 generates the point-to-point response content at the top of each feature point (default is on).
key_flag=1 adds the importance level of each requirement to the proposal headings (default is on).
last_heading_1=2 specifies the starting chapter for the technical solution in “proposal_content.docx.” 8. Run Generate.py. Summary With this tool, you only need to check the product features against the proposal requirements, and most of the proposal content is generated automatically. The business proposal section can also be generated similarly, so creating a 1,000-page proposal now only takes a few minutes. Recently, I haven’t been updating my open-source articles as frequently — not because I’ve stopped writing, but because the progress on open-source commercialization has been great, and the endless task of drafting proposals has consumed my time. As a native open-source commercial company, WhaleOps employs mostly engineers. Asking these open-source contributors to write proposals wastes their development time, and their proposals don’t always meet the quality standard. Unlike managing in a big company, being a startup CEO means stepping into any role the company needs. After every strategic meeting, I’m the first to roll up my sleeves and tackle the most urgent tasks. As a result, I haven’t had time to write articles, as my limited time is mostly taken up with creating proposals that follow the formal template style. Especially recently, with one bid after another, I’ve found myself questioning my sanity. The existing large models can’t solve my problem, so I’ll build my own As a tech person, I always look to tools to solve my problems. Naturally, I thought of large models, but they can’t fully grasp the nuances of our products and often produce unreliable outputs. Plus, you’re the one delivering the work, not the model. So, I decided to develop a tool for proposal generation based on large models using Python and a compatible large model (compatible with ChatGPT). This tool automatically generates a proposal from your product documentation, breaking down product manuals into feature points. Based on a human-created mapping of these points to the requirements, it generates a Word version of the proposal and an Excel deviation table. The model can condense or expand content or simply copy relevant functionality as needed. Features of the open-source proposal tool The functionality of this tool is simple, with the most challenging part being the Word formatting (formatting in Word is always a pain). I experimented with several methods to make sure it follows the correct Title 1, Title 2, Title 3, body text, table, and image formats in the Word template. Staying true to the open-source spirit, I’ve uploaded the tool to my personal GitHub under an Apache License, so feel free to use it if you need. Here’s what it does: Get it here: https://github.com/William-GuoWei/ProposalLLM Get it here: https://github.com/William-GuoWei/ProposalLLM https://github.com/William-GuoWei/ProposalLLM Breaks down your product manual into a set of reusable detail documents, reducing the need to repeatedly reference the source document when drafting a proposal. You can also customize specific functionality (the default file name is “Template.docx”). Based on the requirements table filled in by a person, it automatically generates a proposal in a point-to-point response format, including all headings and content, with the correct Title 1, 2, 3 formatting, and automatically organizes body text, images, and bullet points (the default requirements table is “requirements_table.xlsx,” and the generated content is in “proposal_content.docx”). For any product requirements in the corresponding functionality section, it automatically copies the product manual content into the point-to-point response section, retaining images, tables, and bullet points. You can also rewrite the product description to suit different proposal needs. If there’s no matching functionality, the model automatically generates relevant content (review and modify as needed). Completes the technical requirements deviation table by automatically filling in responses in the “requirements_table.xlsx,” with responses formatted as “Answer: Fully supports, {Model-generated text based on project requirements}” and includes the section number corresponding to the proposal. Breaks down your product manual into a set of reusable detail documents, reducing the need to repeatedly reference the source document when drafting a proposal. You can also customize specific functionality (the default file name is “Template.docx”). Based on the requirements table filled in by a person, it automatically generates a proposal in a point-to-point response format, including all headings and content, with the correct Title 1, 2, 3 formatting, and automatically organizes body text, images, and bullet points (the default requirements table is “requirements_table.xlsx,” and the generated content is in “proposal_content.docx”). For any product requirements in the corresponding functionality section, it automatically copies the product manual content into the point-to-point response section, retaining images, tables, and bullet points. You can also rewrite the product description to suit different proposal needs. If there’s no matching functionality, the model automatically generates relevant content (review and modify as needed). Completes the technical requirements deviation table by automatically filling in responses in the “requirements_table.xlsx,” with responses formatted as “Answer: Fully supports, {Model-generated text based on project requirements}” and includes the section number corresponding to the proposal. With this tool, you can quickly modify and generate proposals at will. Proposal generation process Start by running Extract_Word.py to generate your product feature points as a Word document, then run Generate.py . If a feature point is missing, you can enter “X” in the Excel file, and the model will generate content that meets the requirements. However, I strongly recommend a manual review. After generation, you’ll see the proposal formatted with a table of contents, body text, images, tables, and bullet points, all automatically organized. Extract_Word.py Generate.py The entire proposal is structured according to the client’s requirements in sequential format, with subheadings, content, images, and point-to-point responses. This takes care of all the repetitive work. The automatically generated deviation table includes everything, whether the content was generated by the model or not, along with the corresponding section numbers. You only need to finalize the deviation table with “&” symbols based on the final requirements — the model has written it all for you. For me, this tool has reduced what used to be 8 hours of work to around 30 minutes, and even our business team can generate the business proposal sections using the template. In total, this has cut down our time for a week-long proposal to 1–2 days, reducing the manpower required by 80%. How to use it? First, download all the code to a single directory from GitHub at https://github.com/William-GuoWei/ProposalLLM : Proposal Large Model (Chinese version). https://github.com/William-GuoWei/ProposalLLM Install the Python environment and packages: pip install openpyxl, docx, openai, requests, docx, python-docx. Apply for a ChatGPT or Baidu Qianfan large model key (I used ERNIE-Speed-8K, which is free), record the token, and place it in the relevant code section. Copy your product manual to Template.docx. Be sure to use the body text, Title 1, Title 2, and Title 3 styles provided; other formats may cause issues. Run Extract_Word.py to generate the feature point document from the product manual (supports up to 3 heading levels). If the list formatting appears off, don’t worry; the final format will align properly. Fill in Columns B and C (which will generate secondary and tertiary headings automatically) and Column G (the corresponding product manual chapter). If a chapter is missing, enter “X.” Note that if there is no corresponding chapter or an “X” is entered, the model will generate content automatically. Review the “proposal_content.docx” document and keep the chapter for which you want to start generating the proposal. You can modify the body text and heading 1, 2, and 3 styles; just don’t rename the styles, or there may be errors. Adjust the parameters in Generate.py: Install the Python environment and packages: pip install openpyxl, docx, openai, requests, docx, python-docx . pip install openpyxl, docx, openai, requests, docx, python-docx Apply for a ChatGPT or Baidu Qianfan large model key (I used ERNIE-Speed-8K, which is free), record the token, and place it in the relevant code section. Copy your product manual to Template.docx . Be sure to use the body text, Title 1, Title 2, and Title 3 styles provided; other formats may cause issues. Template.docx Run Extract_Word.py to generate the feature point document from the product manual (supports up to 3 heading levels). If the list formatting appears off, don’t worry; the final format will align properly. Extract_Word.py Fill in Columns B and C (which will generate secondary and tertiary headings automatically) and Column G (the corresponding product manual chapter). If a chapter is missing, enter “X.” Note that if there is no corresponding chapter or an “X” is entered, the model will generate content automatically. Review the “proposal_content.docx” document and keep the chapter for which you want to start generating the proposal. You can modify the body text and heading 1, 2, and 3 styles; just don’t rename the styles, or there may be errors. Adjust the parameters in Generate.py : Generate.py API_KEY and SECRET_KEY: Baidu Cloud large model keys. MAX_WIDTH_CM: Maximum image width; images larger than this will be resized automatically. The prompts for generating content have been customized for large data scenarios, but feel free to modify them. MoreSection=1 will read Column C to generate detailed tertiary headings (default is on). ReGenerateText=0 will re-generate text content automatically for different proposal needs (default is off). DDDAnswer=1 generates the point-to-point response content at the top of each feature point (default is on). key_flag=1 adds the importance level of each requirement to the proposal headings (default is on). last_heading_1=2 specifies the starting chapter for the technical solution in “proposal_content.docx.” API_KEY and SECRET_KEY : Baidu Cloud large model keys. API_KEY SECRET_KEY MAX_WIDTH_CM : Maximum image width; images larger than this will be resized automatically. MAX_WIDTH_CM The prompts for generating content have been customized for large data scenarios, but feel free to modify them. MoreSection=1 will read Column C to generate detailed tertiary headings (default is on). MoreSection=1 ReGenerateText=0 will re-generate text content automatically for different proposal needs (default is off). ReGenerateText=0 DDDAnswer=1 generates the point-to-point response content at the top of each feature point (default is on). DDDAnswer=1 key_flag=1 adds the importance level of each requirement to the proposal headings (default is on). key_flag=1 last_heading_1=2 specifies the starting chapter for the technical solution in “proposal_content.docx.” last_heading_1=2 8. Run Generate.py . Generate.py Summary With this tool, you only need to check the product features against the proposal requirements, and most of the proposal content is generated automatically. The business proposal section can also be generated similarly, so creating a 1,000-page proposal now only takes a few minutes.

I Built An Automatic Proposal Generation Large Language Model and Open-Sourced It on GitHub

MLOps With Databricks and Spark - Part 1

Getting Started with Data Analytics in Python Using PyArrow

Lua is a highly versatile programming language. It is widely used in gaming, embedded systems, web development, networking, automation, and customization. Its small size, flexibility, and speed make it a great choice for embedding into larger applications or for use in environments with constrained resources. When using Lua, we can’t avoid tables. Lua tables are flexible and can be used as arrays, dictionaries (hash maps), or even mixed structures. Tables have no fixed size and can grow based on our needs. Meanwhile, there is one important thing we should not forget: ordering in tables. Order insertion in tables Lua tables, when used as arrays, maintain the insertion order, which makes them a good candidate for lists or arrays in various use cases, including with Redis. In this case, tables can be created manually in the script or returned as a result by third-party libraries, like Redis. Let’s create the table manually and check the insertion order: steps = {}
for i = 1, 10, 1 do
    table.insert(steps,i)
end After the loop, steps = {1; 2; 3; 4; 5; 6; 7; 8; 9; 10} Let’s check what happens when Redis returns a table as a result. redis.call("HSET", 'users', 345, "John", 122, "Bob", 543, "Sam")
local dataDb = redis.call("HGETALL", 'users') After the script is executed, dataDb = {"345", "John", "122", "Bob", "543", "Sam"} In both scenarios, we can find the created table as an array where the insertion order is preserved. But what if we need to convert the Redis response (which is a table as an array) into a key-value associative array? For the first example, we store the following key-value pairs in Redis and want to use the same structure in Lua when processing: {
     345: {name: “John”},
     122: {name: “Bob”},
     543: {name: “Sam”},
     202: {name: “Peter”}
} Run the following script and check the created associative array: redis.call("HSET", 'users', 345, "John", 122, "Bob", 543, "Sam", 202, "Peter")
 
local dataDb = redis.call("HGETALL", 'users')
 
local directory = {}
 
for i = 1, #dataDb, 2 do
    directory[dataDb[i]] = dataDb[i+1]
end where, after the loop dataDb = {"345"; "John"; "122"; "Bob"; "543"; "Sam"; "202"; "Peter"} directory = {["122"]="Bob"; ["345"]="John"; ["202"]="Peter"; ["543"]="Sam"} Let's look at another example of using the cjson module: redis.call('SET', 'packages', '{"10":{"name":"Plus"},"15":{"name":"Family"},"25":{"name":"Premium"}}')
 
local pkgsDb = redis.call('GET', 'packages')
local parsedPkgs = cjson.decode(pkgsDb) where: pkgsDb = "{\"10\":{\"name\":\"Plus\"},\"15\":{\"name\":\"Family\"},\"25\":{\"name\":\"Premium\"}}" parsedPkgs = {["25"]={["name"]="Premium"}; ["15"]={["name"]="Family"}; ["10"]={["name"]="Plus"}} It is good to remember that the CJSON module returns a table in a different order than the one parsed in the JSON. This can prevent errors and save you time, as debugging such problems is quite difficult. In both scenarios, we may find that the insertion order is not guaranteed in a table as an associative array. How to sort a table key when needed? Sometimes, having ordered tables in Lua is important. The order of elements can be critical for the functionality that the script provides. We can use the built-in table.sort function to sort tables. It receives the array to be sorted, along with an optional order function. If this function is not provided, sort uses the default mode. All numeric values will be sorted in ascending order, and all strings will be sorted alphabetically. Sorting table – array In the case of simple sorting of numeric or string values, it is enough to use the following expression: local num = {4, 2, 6}
local str = {"pear" ,"apple", "melon"}
 
table.sort(num)
table.sort(str) where sorted  num = {2; 4; 6} and str = {"apple"; "melon"; "pear"} If the sorting is based on a certain condition, we can define a comparison function. local t = {4,7,2,3}
table.sort(t, function(a, b) return a > b end) where sorted   t = {4; 7; 2; 3} Sorting a table – dictionary To sort a table structured as a dictionary, we can sort it by key or value.
This requires several steps: Extract the keys or values into an array-like table


Sort this array


Iterate over the sorted keys and get the value from the original array by key Sort by keys: local transport = {[100]="train", [70]="tram", [90]="bus"}
 
local keys = {}
for key in pairs(transport) do
    table.insert(keys, key)
end
 
table.sort(keys)
 
for _, key in ipairs(keys) do
    print(key, transport[key])
end Script result: 70 tram, 90 bus, 100 train Sort by values: local entities = {}
 
for key, value in pairs(transport) do
    table.insert(entities, {key = key, value = value})
end
 
table.sort(entities, function(a, b) return a.value < b.value end)
 
for _, entity in ipairs(entities) do
    print(entity.key, entity.value)
end Script result: 90 bus, 100 train, 70 tram Conclusion Having ordered tables in Lua is important when the order of elements matters. In Lua, tables that function like dictionaries have no inherent order, meaning that when we iterate over them using pairs(), the order of elements is not guaranteed. In this article, we explored how to sort arrays in Lua using table.sort and how to manage separate key and value arrays for sorting dictionaries. Lua is a highly versatile programming language. It is widely used in gaming, embedded systems, web development, networking, automation, and customization. Its small size, flexibility, and speed make it a great choice for embedding into larger applications or for use in environments with constrained resources. When using Lua, we can’t avoid tables. Lua tables are flexible and can be used as arrays, dictionaries (hash maps), or even mixed structures. Tables have no fixed size and can grow based on our needs. Meanwhile, there is one important thing we should not forget: ordering in tables . ordering in tables Order insertion in tables Order insertion in tables Lua tables, when used as arrays, maintain the insertion order, which makes them a good candidate for lists or arrays in various use cases, including with Redis . Redis In this case, tables can be created manually in the script or returned as a result by third-party libraries, like Redis. Let’s create the table manually and check the insertion order: steps = {}
for i = 1, 10, 1 do
    table.insert(steps,i)
end steps = {}
for i = 1, 10, 1 do
    table.insert(steps,i)
end After the loop, steps = {1; 2; 3; 4; 5; 6; 7; 8; 9; 10} After the loop, steps = {1; 2; 3; 4; 5; 6; 7; 8; 9; 10} Let’s check what happens when Redis returns a table as a result. redis.call("HSET", 'users', 345, "John", 122, "Bob", 543, "Sam")
local dataDb = redis.call("HGETALL", 'users') redis.call("HSET", 'users', 345, "John", 122, "Bob", 543, "Sam")
local dataDb = redis.call("HGETALL", 'users') After the script is executed, dataDb = {"345", "John", "122", "Bob", "543", "Sam"} After the script is executed, dataDb = {"345", "John", "122", "Bob", "543", "Sam"} In both scenarios, we can find the created table as an array where the insertion order is preserved. In both scenarios, we can find the created table as an array where the insertion order is preserved. But what if we need to convert the Redis response (which is a table as an array) into a key-value associative array? For the first example, we store the following key-value pairs in Redis and want to use the same structure in Lua when processing: {
     345: {name: “John”},
     122: {name: “Bob”},
     543: {name: “Sam”},
     202: {name: “Peter”}
} {
     345: {name: “John”},
     122: {name: “Bob”},
     543: {name: “Sam”},
     202: {name: “Peter”}
} Run the following script and check the created associative array: redis.call("HSET", 'users', 345, "John", 122, "Bob", 543, "Sam", 202, "Peter")
 
local dataDb = redis.call("HGETALL", 'users')
 
local directory = {}
 
for i = 1, #dataDb, 2 do
    directory[dataDb[i]] = dataDb[i+1]
end redis.call("HSET", 'users', 345, "John", 122, "Bob", 543, "Sam", 202, "Peter")
 
local dataDb = redis.call("HGETALL", 'users')
 
local directory = {}
 
for i = 1, #dataDb, 2 do
    directory[dataDb[i]] = dataDb[i+1]
end where, after the loop where, after the loop dataDb = {"345"; "John"; "122"; "Bob"; "543"; "Sam"; "202"; "Peter"} dataDb = {"345"; "John"; "122"; "Bob"; "543"; "Sam"; "202"; "Peter"} directory = {["122"]="Bob"; ["345"]="John"; ["202"]="Peter"; ["543"]="Sam"} directory = {["122"]="Bob"; ["345"]="John"; ["202"]="Peter"; ["543"]="Sam"} Let's look at another example of using the cjson module: redis.call('SET', 'packages', '{"10":{"name":"Plus"},"15":{"name":"Family"},"25":{"name":"Premium"}}')
 
local pkgsDb = redis.call('GET', 'packages')
local parsedPkgs = cjson.decode(pkgsDb) redis.call('SET', 'packages', '{"10":{"name":"Plus"},"15":{"name":"Family"},"25":{"name":"Premium"}}')
 
local pkgsDb = redis.call('GET', 'packages')
local parsedPkgs = cjson.decode(pkgsDb) where: where: pkgsDb = "{\"10\":{\"name\":\"Plus\"},\"15\":{\"name\":\"Family\"},\"25\":{\"name\":\"Premium\"}}" pkgsDb = "{\"10\":{\"name\":\"Plus\"},\"15\":{\"name\":\"Family\"},\"25\":{\"name\":\"Premium\"}}" parsedPkgs = {["25"]={["name"]="Premium"}; ["15"]={["name"]="Family"}; ["10"]={["name"]="Plus"}} parsedPkgs = {["25"]={["name"]="Premium"}; ["15"]={["name"]="Family"}; ["10"]={["name"]="Plus"}} It is good to remember that the CJSON module returns a table in a different order than the one parsed in the JSON. This can prevent errors and save you time, as debugging such problems is quite difficult. It is good to remember that the CJSON module returns a table in a different order than the one parsed in the JSON. This can prevent errors and save you time, as debugging such problems is quite difficult. In both scenarios, we may find that the insertion order is not guaranteed in a table as an associative array. In both scenarios, we may find that the insertion order is not guaranteed in a table as an associative array. How to sort a table key when needed? How to sort a table key when needed? Sometimes, having ordered tables in Lua is important. The order of elements can be critical for the functionality that the script provides. We can use the built-in table.sort function to sort tables. It receives the array to be sorted, along with an optional order function. If this function is not provided, sort uses the default mode. All numeric values will be sorted in ascending order, and all strings will be sorted alphabetically. table.sort Sorting table – array Sorting table – array In the case of simple sorting of numeric or string values, it is enough to use the following expression: In the case of simple sorting of numeric or string values, it is enough to use the following expression: local num = {4, 2, 6}
local str = {"pear" ,"apple", "melon"}
 
table.sort(num)
table.sort(str) local num = {4, 2, 6}
local str = {"pear" ,"apple", "melon"}
 
table.sort(num)
table.sort(str) where sorted  num = {2; 4; 6} and str = {"apple"; "melon"; "pear"} where sorted  num = {2; 4; 6} and str = {"apple"; "melon"; "pear"} If the sorting is based on a certain condition, we can define a comparison function. local t = {4,7,2,3}
table.sort(t, function(a, b) return a > b end) local t = {4,7,2,3}
table.sort(t, function(a, b) return a > b end) where sorted   t = {4; 7; 2; 3} where sorted   t = {4; 7; 2; 3} Sorting a table – dictionary Sorting a table – dictionary To sort a table structured as a dictionary, we can sort it by key or value.
This requires several steps: Extract the keys or values into an array-like table Sort this array Iterate over the sorted keys and get the value from the original array by key Extract the keys or values into an array-like table Extract the keys or values into an array-like table Sort this array Sort this array Iterate over the sorted keys and get the value from the original array by key Iterate over the sorted keys and get the value from the original array by key Sort by keys: Sort by keys: local transport = {[100]="train", [70]="tram", [90]="bus"}
 
local keys = {}
for key in pairs(transport) do
    table.insert(keys, key)
end
 
table.sort(keys)
 
for _, key in ipairs(keys) do
    print(key, transport[key])
end local transport = {[100]="train", [70]="tram", [90]="bus"}
 
local keys = {}
for key in pairs(transport) do
    table.insert(keys, key)
end
 
table.sort(keys)
 
for _, key in ipairs(keys) do
    print(key, transport[key])
end Script result: 70 tram, 90 bus, 100 train Script result: 70 tram, 90 bus, 100 train Sort by values: Sort by values: local entities = {}
 
for key, value in pairs(transport) do
    table.insert(entities, {key = key, value = value})
end
 
table.sort(entities, function(a, b) return a.value < b.value end)
 
for _, entity in ipairs(entities) do
    print(entity.key, entity.value)
end local entities = {}
 
for key, value in pairs(transport) do
    table.insert(entities, {key = key, value = value})
end
 
table.sort(entities, function(a, b) return a.value < b.value end)
 
for _, entity in ipairs(entities) do
    print(entity.key, entity.value)
end Script result: 90 bus, 100 train, 70 tram Script result: 90 bus, 100 train, 70 tram Conclusion Having ordered tables in Lua is important when the order of elements matters. In Lua, tables that function like dictionaries have no inherent order, meaning that when we iterate over them using pairs() , the order of elements is not guaranteed. In this article, we explored how to sort arrays in Lua using table.sort and how to manage separate key and value arrays for sorting dictionaries. pairs() table.sort

Stories, News & Mentions about TABLEXI

Tablexi’s Stories on HackerNoon

Tablexi's News & Mentions