Hackernoon logoBuilding AI-Based Trading Algorithms - An Interview with Kiran Mannam, CEO at Rocket Vault Finance by@Ishan Pandey

Building AI-Based Trading Algorithms - An Interview with Kiran Mannam, CEO at Rocket Vault Finance

image
Ishan Pandey Hacker Noon profile picture

@Ishan PandeyIshan Pandey

Student of law working on code and everything law. Founder: Blockchain Research

Ishan Pandey: Hi Kiran, welcome to our series “Behind the Startup.” Please tell us about yourself and the story behind Rocket Vault Finance?

Kiran Mannam: Hey Ishan, thanks for having me over. Delighted to be on Hackernoon and share more details about the exciting stuff we are developing at Rocket Vault.

I come from the Financial Services domain and have spent more than 17 years working with various financial institutions across the globe, understanding their complex problems and providing solutions to address those.

The average retail investor entering the Crypto/DeFi landscape is confused. There are more than 8,000 different cryptocurrencies on the market. Following the market, development takes too much time and effort when done manually. By the time one comes to a conclusion, the best opportunity is already lost.

Rocket Vault started with a simple idea to simplify the value of investing in crypto assets for the common person. We wanted to develop a product that is really simple to use from an end-user perspective and handles all the complex parts behind the scenes that minimises risks and maximises returns and most importantly, lets you be at peace while Rocket Vault works hard for you 24x7x365 days.

Ishan Pandey: ‘Stablecoins’ are crypto-assets that aim to maintain a stable value relative to a specific asset or a pool or basket of assets. What role does Machine Learning Algorithms play in minimising losses and maximising gains?

Kiran Mannam: The best part about Rocket Vault trading system is that it adopts a data-driven approach and keeps the “emotion” quotient away.

The trading-based AI engine is fed with data from global exchanges and it applies data modelling techniques and statistical analytics to analyse the data. Using advanced AI predictive analytics & Machine Learning pattern recognition algorithms identifies the risky assets and avoids them and thereby minimises risks.

It also identifies the best-performing assets and diversifies the portfolio across multiple high performing assets ranging from 500-800, which is not possible to achieve manually. The algorithm also keeps monitoring the investments in real-time to keep booking profits from time-to-time at optimal levels. Another advantage is that the profits generated are compounded automatically, thereby maximising gains.

Ishan Pandey: Can you explain how Smart Vaults work and how these strategies help increase the return rate? (Technical explanation, code, trading modules, front-end etc.)

Kiran Mannam: Our Smart Vault grabs the various short term opportunities using AL/ML trained models using the Combination of Anomalies in volume and time-series data patterns irrespective of bull or bear market conditions.

If you notice our LIVE order alerts in the telegram channel, for every Sell order, the time difference between (BUY Low & SELL High) or (Sell High & Buy Low) varies between minutes, hours or max days. Rarely you will find a prediction that lasts weeks or months.

Our trading algorithm is not designed to predict an asset’s price moment for a very long-time frame, like months or years. Each prediction targets smaller profits with a high probability of being truly positive and with a limited risk of being a false positive.

The predictions work on a probability theory where we are counting on true overall prediction with smaller profits in a short time frame and more accumulated profit in a long period of time which is a unique competitive advantage of our product.

Using a short term time frame prediction strategy along with a balanced asset allocation strategy, the algorithm eliminates the risks involved irrespective of bull or bear market conditions in the long run. This is the reason why the payout structure is also quarterly.

Currently, the true prediction rate stands at around 85% and as the AI engine keeps learning from data, it keeps getting smarter and leads to more accurate predictions in the future.

Ishan Pandey: I read that you have been developing, testing & executing “Smart Vault” strategies which essentially replace an experienced fund manager by building intelligence into the trading system. Do you think that AI traders are going to rule the market? Further, what does this mean for retail investors who don’t have the resources to build AI-based trading algorithms?

Kiran Mannam: You have rightly pointed out that many retail investors don’t have sufficient resources/infrastructure to capture all the market opportunities by using the latest technology available out there. Our motivation is to create a pool for retail investors, which we would offer a free service just by holding RVF tokens and trading algorithm will take care of grabbing all such opportunities and sharing the profits back to pools.

AI traders already impacting the financial world, be it Equities, Fx or Bond Markets and Crypto would be no different.

Regarding replacing a fund manager, generally fund managers are equipped with a lot of data that would help them decide how much to invest, book profit/disinvestment, etc.

All these decisions are taken manually with the help of data and a lot of experience gathered over a period of time.

We follow a similar approach. All decisions are data-driven, backed by historical data and completely automated in making value investment in the long term. We need to feed various data sources to AI engines for a better decision-making process. I believe we will see many advancements in value investments going forward in the following years.

Ishan Pandey: As opposed to trading bots that require a significant amount of user input and decision-making, Rocket Vaults are automated smart vaults that hold cryptocurrencies in a pool, applying artificial intelligence and machine learning algorithms. How safe are these smart vaults in terms of Client protection and what is the technology framework?

Kiran Mannam: We have been developing the Vault Strategies since early 2019 and actual testing started from late 2019 - early 2020. It has gone through many testing phases. All these times while we are developing our product, we have incorporated an effective feedback mechanism that controls vault’s behaviour in extreme scenarios using separate control engines, which we refer to as “Treasury Management Engine”, “Asset Rebalancing Engine”, “Monitoring Engine” among others that monitors the Vaults to handle all possible scenarios.

I recommend taking a look at the white paper on our website for an in-depth understanding of different engines’ role.

Ishan Pandey: What kind of impact do you think will The Stablecoin Tethering and Bank Licensing Enforcement (STABLE) Act have on Stabelcoins in the US Market?

Kiran Mannam: I believe the STABLE act would significantly increase regulation and supervision of stablecoins, treating them more like traditional brick-and-mortar banks under U.S. law. I believe some level of regulation is good for the greater adoption of digital currencies. We have been watching the space closely and being compliant is a top priority for us.

The trading algorithm is not limited to stablecoins vaults as such. We do have plans to launch ETH, BTC and other vaults too going forward. The user is free to choose which Vault services he would like to avail.

Whichever token the user deposits, the rewards are distributed in the same token.

Ishan Pandey: What factors can affect the APY of an algorithm? Further, when coding trading algorithms, what things should developers keep in mind?

Kiran Mannam: Well, investors love the crypto space because of one main reason: “high volatile behaviour”. If the whole crypto market stays calm, there is not much scope to predict the vault, impacting the APY. So market volatility is one of the factors that decide the APY.

Our main focus while developing AL/ML strategies is primarily to improve the predictability based on historical data while adapting to the current market conditions. And we never rely on a single strategy that could potentially be outdated with new market conditions. Keep adjusting the AI training models and adjust the hyperparameters that reflect the current market conditions.

It is recommended to have some risk management in place, alerting the right team when the system sees red flags.

Ishan Pandey: According to you, what new trends are we going to see in the Stablecoins market moving forward?

Kiran Mannam:

The number of stable coins in circulation surged by 500% in 2020 — cementing their reputation as “one of the most-used payment networks in the world.”

Looking ahead, I believe that the value of dollar-backed stablecoins will surpass $150 billion and also, there could be more stable coins joining the likes of Tether and USDC.

The purpose of this article is to remove informational asymmetry existing today in our digital markets by performing due diligence by asking the right questions and equipping readers with better opinions to make informed decisions. The material does not constitute any investment, financial, or legal advice. Please do your research before investing in any digital assets or tokens, etc. The writer does not have any vested interest in the company.

Tags

Join Hacker Noon

Create your free account to unlock your custom reading experience.