Trading & Risk Systems

Problem
Commodities and futures trading software has almost no tolerance for lag — a hedging dashboard that lags the market or a notification that arrives late is simply wrong.
What we built
Optimised API controllers that cut P90 latency from ~500ms to ~150ms at 60 requests/second; a real-time trading service scaled from one region to 114 countries on SignalR with backpressure handling; a decoupled notification engine on SignalR and Azure Communication Services.
Impact
A 70% cut in P90 latency, a global rollout with zero downtime, and a fault-tolerant notification and reporting layer kept off the critical path of the main API.
Technology
- .NET Core
- React
- TypeScript
- SignalR
- Azure
- MSSQL
- Elasticsearch