Job Description
What you'll build
You will own services where milliseconds matter and correctness is non-negotiable. Order routing across CEX, DEX, OTC venues. Execution engine that aggregates liquidity and chooses best fill in real time. Position management. Risk checks that fire before order leaves perimeter. Market data ingestion at exchange-connectivity level. Code you write touches client capital directly - every flag, every retry, every timeout has financial consequence.
This is closest team to actual market structure inside Mainframe. You will read venue API docs more often than you read Slack. You will care about WebSocket reconnection logic the way other engineers care about CSS. You will find latency you did not know was there.
Who fits
You have shipped systems where trade did not execute means revenue did not happen. Distributed systems under real-time pressure are home territory. You think in event streams, not request-response. You have argued about exactly-once semantics and have opinions. Exchange connectivity, FIX protocol, market microstructure, MEV - bonus across the board. People who built matching engines, prime brokerage backends, HFT infrastructure are exactly who we want.
Stack signals:
Go or Rust at production depth (we use both - pick your weapon)
gRPC + Protobuf for service-to-service, WebSocket for venue connectivity
PostgreSQL + Redis comfort under high write throughput
Kubernetes, Docker, AWS at production level
Centrifugo or similar real-time messaging layer experience welcome
