New Delhi, Aug 09, 2022, IANS

New Delhi, Aug 9 (IANS) As trading in the stock market becomes more data and analytics-driven, the challenge for retail traders -- both the new and seasoned ones -- has always been what stock to buy or sell and when to buy or sell (exact time).

The number of active Demat accounts in the country jumped 63 per cent to 89.7 million in the financial year 2021-22 (FY22), riding on high smartphone usage, easier digital onboarding of customers and attractive returns delivered by the equity markets.

In a dynamic and ever-evolving stock investment scenario today, AlgoBulls which was founded in 2019 by Pushpak Dagade, Jimmit Patel and Suraj Bathija, provides automated and customised artificial intelligence (AI)-based experiences of trading in stock markets to everyone.

Dagade, Founder and CEO, AlgoBulls, tells IANS that the platform has built a Deep Tech stack for complete end-to-end automation for retail traders, with deep integrations with multiple brokerage firms and market experts to convert their trading strategies into algorithms.

Here are the excerpts from an interview:

Q: When did you first get interested in the markets? What made you take the entrepreneurial leap and start AlgoBulls?

A: After graduating from IIT Delhi, I began working in Bengaluru, where I was first introduced to trading by my colleagues. But eventually, when it came to the volatile world of trading, tips from friends, family, and news channels did more harm than good, so I began my own research to understand how a layman can begin trading.

To bring a solution to the table, one must first understand the problem. Back then, the top broking houses just started coming up with retail trading APIs. This meant anyone with a Demat Account could place a trade on the broker via trading APIs. But, in order to reach an end-to-end complete solution, there were numerous challenges involved.

The alarming issues in the existing ecosystem were the lack of trading knowledge -- what (stock) to buy/sell & when to buy/sell (exact time), the inability to grab good market opportunities in near real-time & trading with emotions.

For me, this was the inspiration behind demonstrating a simple trading strategy using Python. I created a strategy that could fetch live market data in real-time, use it to compute signals and execute a trade when there was a buy or sell signal - trading with complete automation.

This strategy was sent as a proposal to PyCON India, where I received a great response from other traders and investors wanting to switch to automated trading. This gave birth to my vision to create -- AlgoBulls, along with my two friends, Suraj Bathija and Jimmit Patel.

We have built a Deep Tech stack for complete end-to-end automation for retail traders. We have deep integrations with multiple brokerage firms and have collaborated with market experts to convert their trading strategies into algos. AlgoBulls has gained prominence in the trading industry in a relatively short period of time.

Q: Tell us about your long-term vision and mission for AlgoBulls.

A: Our vision is to bring financial inclusion to adults of all ages by offering a platter of financial products for various returns and risks, such as high risk -- high returns, medium risk -- medium returns, and low risk -- good returns, which are backed by technology and automation.

Currently, we are on a mission to democratise algo trading for retailers and enable access to alpha-generating products a reality for everyone. In contrast to HNIs or institutions, AlgoBulls intends to provide equal opportunities for retail investors.

Q: What differentiates you from other entities providing AI-driven algo strategies?

A: At AlgoBulls, we have built India's First Trading Marketplace. Deep Tech stack, fully automated trade execution, transparent and authentic trading marketplace, and getting access to an expert trader's decades worth of knowledge in just a few clicks are some of the things which greatly differentiate us from our competitors.

For regulatory reasons, anyone who wants to stay in the market must obtain the algo IDs for each retail algo strategy they run in the near future. Their platform & strategy should be approved before they can procure these algo IDs. In the coming months, regulatory approvals at multiple levels will be required. This is one of the major places where we have a significant advantage over other companies.

The strategies listed on AlgoBulls Marketplace have been thoroughly vetted, authorised, and deployed in accordance with a strict procedure.

The strategy performance data is updated daily on the website based on the day's live trading data and returns are updated based on rolling window periods. Transparency and authenticity of strategy performance data are key to getting customers' confidence in the quality of trading strategies. No other player is doing so currently.

To execute trades in seconds, technology and data management are the most critical tasks that are handled extremely efficiently, thanks to our platform's highest standard of technology and a Deep Tech stack.

Q: There has been a huge surge in retail participation in financial markets post the pandemic. How do you educate them when it comes to complex trading strategies?

A: Post pandemic, many beginner traders have dipped their toes into the volatile waters of day trading with little or no knowledge. The ever-growing appetite for shorter-term returns has led a huge chunk of Millennial and Gen Z investors to dive into the retail markets to compete with the big sharks.

To ensure such participants have access to trading knowledge, we have collaborated with over 30+ brokerage firms, numerous traders, and multiple strategists to build a full ecosystem that allows the trading community to generate alpha in tough markets.

AlgoBulls hosts a number of webinars with major brokerage firms, as well as on-site events and tutorials for retail investors. To ensure that our users understand what our algorithmic strategies offer, the webinars held by our market expert strategists are absolutely free.

In addition, we have a digital talk show called 'Traders Talk'. Hosted by one of our co-founders Bathija, this talk show is designed to motivate and inspire traders of all levels.

Our customer support team is always available for customers who want to learn more about Algo Trading.

Q: What has cloud technology and AWS enabled you to do better?

A: We have been running our business on AWS since day 1, and that has paid off very well. AlgoBulls is the first platform to provide a fully automated trading platform for retail traders in India.

Initially, we started with a monolith application but as our usage grew, we broke down our monolith into microservices. Today we are running most of our services using serverless architecture. This has helped us reduce our costs by nearly 60 per cent as compared to the traditional approach while helping us easily achieve 10x scale and 2x better performance with no additional efforts on the scaling aspect.

We use Amazon S3, which helps store tons of data at very reduced costs, for data ingestion and this data after extraction and post-processing is eventually stored in Amazon DynamoDB. This data is then used for ML training. We use ML for generating profitable quant strategies from a pool of ;base strategies'.

We use Amazon ECS with AWS Fargate for executing our client strategies in real-time by leveraging containerization technology.

This means every client's each strategy execution gets dedicated compute, network and memory resources and no two strategies compete with each other for the same. This service helps us save costs greatly while ensuring that even if some strategy is computationally heavy, the other strategy would not slow down as it's not sharing its hardware resources with the others.

We use AWS CloudFormation to easily deploy, upgrade or modify our complex infra for real-time services in a matter of minutes & with reliability. We use GitHub actions to automate our deployments.

Our entire infra is serverless. This means all our resources are used only on a need basis. When there is no demand, these services scale down and they scale up automatically when the number of requests goes up.

Also, our load during NSE market hours (9 a.m. to 3.30 p.m. IST) is quite high as compared to the rest of the day. AWS helps us scale our service capacity based on time schedule. Also, we turn off our real-time data service post-midnight, after the MCX exchange shuts off at around 11.30 p.m., as these services are meaningful only when an exchange is running to fetch live data.

Since a lot of our services are for real-time applications, we use AWS Network Load Balancer (NLB) extensively to move data internally among our services at sub-millisecond latency. Our architectural backbone's data fabric is powered by NLBs for efficient and low latency data transfer for real-time applications.

We have planned our architectures to have deployments of critical services across multiple availability zones, which has helped us have no downtimes so far.