Automated Stock Trading Platform

Target audience

The platform is designed for multiple client segments – from individual investors and traders seeking to diversify their portfolios or generate additional returns by investing in an algorithmic trading strategy, to institutional investors such as hedge funds, venture capital funds, and large investment firms looking to invest a part of their portfolios in an active trading vehicle with low or negative correlation to the rest of their portfolios.

Solution overview

The solution is an algorithmic trading system based on a complex, multi-level analysis of prices and the behavior of their derived characteristics. The system allows the administrator to set up trading strategies with different market instruments and backtest them with data obtained from different financial markets and time frames. The trading strategies continuously analyze the behavior of the indicators and initiate the execution of orders once the predetermined conditions are met. In addition, the system allows for parameters to be fine-tuned and displays calculations and events on charts in real-time mode. The solution consists of three subsystems.

1. Admin subsystem

This subsystem allows admins to set up trading strategies and track price movements on the market data chart, as well as configure order execution policy. The data received from the data provider is displayed on the charts in real time, and the calculation and analysis of all indicators are also done in real time. When configuring a trading strategy, the admin sets parameters for RSI, stochastic conditions, candlestick patterns, and many other indicators. The trade management module allows the admin to set multiple parameters to configure the order execution policy:

• entry price 
• how many orders should be sent (multiple trade execution)
• trade size
• exit price (stop-loss management or profit target)

The admin can initiate algo trading with the pre-configured strategy. In this case, the platform performs technical analyses (based on the received market data and counted indicators) and determines triggers. If the predefined conditions for triggers are met, the strategy is executed and the platform places an order. The system allows traders to test and launch their strategies in several different modes. The strategy lifecycle consists of 3 stages: Backtesting, Paper Trading, and Live Trading. Backtesting The backtesting mode allows traders to test the strategy ”back in time” using historical market data. This is the first check performed to understand how the strategy generally performs. Each time a backtest is performed, a trade report is generated to show how profitable the strategy was within a specific time range. Each time the strategy goes through backtesting, traders can make adjustments to make it more profitable. Paper Trading After the strategy has successfully passed the backtesting stage, traders try to launch it in demo mode (i.e., paper trading), which is much closer to real market conditions. Paper trading is performed with market updates in real time, so the only thing simulated is money. Order management is also very close to the conditions of real exchange trading. Live Trading Once the strategy passes the previous two stages, it moves directly to live trading, performed in real time and with real money. Real market conditions can at times be very tough and order execution logic too complicated, leaving only the best strategies to reach this stage. A strategy that has successfully passed this final stage and is able to earn enough money on the real market holds significant value for the trader.

2. Client subsystem

The second subsystem is intended for users. It allows them to register, add money to accounts via the PelicanPay API, withdraw funds, monitor balances and portfolios, and view trade history. Statistics are presented to clients using charts and tables, and users are able to download reports directly. 

3. Corporate subsystem

This part of the system allows the admin to manage users, check the statistics on trades, make money transfers between the accounts of the system, and manage all of the accounts.

Technical solution highlights

Technically, the system consists of several microservices that use event-based communication (message bus). Microservices are asynchronous Python applications that implement a data processing pipeline, i.e. the result from one microservice is used as input data for another (or multiple microservices). Python Socket.IO is used for real-time updates. From the perspective of UI, it is a React application that uses MobX as state management on the front end.

• We chose the Python language and Django framework for backend development and JavaScript, HTML5, and React JS for the frontend part.
• The AWS cloud infrastructure that our team chose to run the platform reduces the time it takes for the web application to load in the end user’s browser, and provides additional monitoring and data protection tools.
• To meet the security requirements of the broker and payment gateway, we implemented role-based access controls, multi-factor authentication, encryption, and other security measures.
• The general architecture of the platform is defined by high performance requirements. Since trading opportunities can appear and disappear in a fraction of a second, the system must process massive volumes of data simultaneously to react to market events in real time.

Development process

We completed the project following the agile development process, with frequent deliveries, full transparency, and close collaboration with our client. Every two weeks, we met with the client to deliver and demonstrate results and report on costs. Frequent demonstrations and feedback collection sessions allowed us to conduct early reviews of all materials and initial software versions and make timely adjustments to functionality. Initially, the client provided only a few design sketches, so we held a series of workshops to define and obtain approval for the technical solution architecture, break down product functionality into sprints, prioritize features, and define detailed requirements. After that, we moved on to the development phase, following the product roadmap. Fourteen months after the start of the project, the MVP version with its core functionality was up and running. Currently, Itexus engineers are developing and implementing additional features.

Results & future plans

Within a year, Itexus’ cross-functional team of experts delivered an MVP version of an automated, high-performing, and secure trading solution that met all of our client’s requirements. The platform is up and running. In the meantime, we continue to maintain and support the existing solution and are working on the advanced features of the platform as well as the mobile versions of the application. Looking for a talent pool to fill the software development gap for your project? Itexus engineers are ready to tackle your idea. Let’s discuss it.