Skip to main content

Forex Feature Generation and Real- time/Historical testing platform

Our platform consists of two implementations of a custom Feature Network Description language that will generate features for State Street's proprietary machine learning algorithms as well as serve as a back-testing and live Forex trading system. Our Python implementation of the FND will allow for historical back-testing with large amounts of data, and our GOLang implementation utilizes a multi threaded design to quickly ingest live data, process the data, and then send buy/sell orders based on a basic trading strategy we developed. Both implementations utilize similar language syntax for easy cross compatibility. As
a whole, the platform serves as a historical back-testing tool for trading strategies that can also deploy successful strategies on the live market.

Team Members: 

Brian Madina

Teodor Matase

Julia Rebello

Rodolfo Rodriguez

State Street