Sigma Stream

Built a system in Rust, Grafana, and Python to process high-frequency market data and capitalise on market inefficiencies.

Initially, our logic was implemented in Python, but we faced performance bottlenecks. We experimented with Go but ultimately chose Rust for its superior concurrency model and zero-cost abstractions, providing the performance and safety we needed. We also incorporated InfluxDB for time-series data storage and Redis for pub/sub message passing.

In a team of five at a startup, concurrency and correctness were critical. I was responsible for building an order book synthesiser, and Rust's ownership model gave us the safety guarantees we needed at the speeds we required.