Technology Stack at Quantify Capital

Sourabh Sisodiya
5 min readJun 26, 2021

Technology is at the core of how we approach trading & we consider ourselves as much a technology company as a trading firm. It is not that we use technology, we live it. Here is a quick, broad, overview of our tech stack.

Django

We use the Django framework for most of our web apps (in-house scanners) as it encourages rapid development & clean, pragmatic design. Django takes care of much of the hassle of web development so we can focus on writing apps and logic without the need to reinvent the wheel. This drastically reduces the time-to-market of our products and services and helps us to test and deploy ideas real quick. Django ORM (Object Relational Mapper) is a “pythonical” way to communicate with our numerous Postgres database nodes to store and retrieve a gigantic amount of data our market data API’s fetches from the exchange in real-time.

Python

Python is extensively used for backtesting as various open-source modules like pandas, NumPy, Talib, Backtrader, FFN, Matplotlib, and many more data analytics & visualization tools are built on top of it. Python helps us in analyzing large datasets in an easy way. Python is our most preferred language as it helps in rapid development and huge online community support adds as the cream to the bagel.

Postgres

Multiple Postgres database nodes are employed to store and retrieve a gigantic amount of data our market data API’s fetches from the exchange in real-time. Data across multiple instruments, expires, and strikes for the last several years is stored in form of billions of rows into our Postgres database cluster. Postgres Database is not only reliable but also scalable, secure, low cost and is one of the most advanced relational databases present in the market.

Machine Learning

As a Quant Firm, we extensively try to find various statistical measures to model the market. Probability is the word of the game here. For any highly probable event, various deployable strategies are formulated which are then backtested religiously on the highest resolution historical data available across instruments of interest. Once a strategy looks promising Machine Learning is used to find unknown correlations, covariances between parameters and optimize the entry and exit conditions that indeed increase the profitability of a system. Various Regression models are feed with the data to find invisible trends and identifying which variables have an impact on the topic of interest. Decision Trees are employed to split data continuously according to a certain parameter. Let’s say you want to predict whether it is profitable to sell a straddle on a day when IV is at 22? Or when there is a big gap up/ gap down (±4%)? It gives an outcome as yes or no. This helps to understand outliers and loss-making days. After an enormous data crunching the best system is automated and deployed.

Normal strategy vs ML optimized Strategy

AWS & DigitalOcean Cloud Services

All our Database instances and Virtual Compute Machines are spread across the digital ocean and AWS cloud service provides. All of our websites, real-time applications, and time-critical systems are deployed on the cloud to not only ensure zero downtime but also scalability, high performance, and low costs.

Scheduling

The software utility cron, also known as a cron job is a time-based job scheduler in Unix-like computer operating systems. Under the hood of our various automated trading algos, alerts systems, risk management systems lie countless cron jobs which help schedule tasks. Software interrupts are employed to manage synchronous events for our Algo trading bots.

Risk Management System

A Risk Management System helps cut down losses. A master node collects all the events & logs them on a dashboard as a centralized knowledge hub for all open positions across the firm. To keep check of an unsystematic risk event like exchange blackout etc the master dashboard provides endpoints to manually manage trades.

This blog is written by Rahul Sanghvi (Associate Software Developer at Quantify Captial), lets look at what he has to say to the readers -

“Are you as Intrigued by technology as I do? Or you want to discuss some amazing stuff around technology and stock markets? Feel free to connect with me on LinkedIn. I Love being around all Sci-Fi stuff as they show in movies. I’m a Full-Stack Developer and Stock Market Enthusiast. To know more about me and my projects please visit my website. I hope I did uncover the behind the scene technology and processes we use that go into delivering a visceral experience to the end-product users. Thanks for reading :)

--

--