Connect HFT Strategy to Broker

I already have a fully-coded high-frequency trading strategy that has been stress-tested in simulation; now it must execute live on stocks and ETFs with the same sub-millisecond performance. What I need is the technical bridge between my code and a broker or exchange that offers true direct market access with optional colocation.

Your task is to design and implement a rock-solid order-routing layer that links my strategy to the venue’s FIX, OUCH or proprietary low-latency gateway, handles real-time market data, and keeps every round-trip as short as possible. Expect to work with C++ (latency-critical sections) and Python (strategy logic), although I’m open to the tools you believe will shave microseconds.

Key deliverables
• Connection module that authenticates, maintains heartbeats and recovers seamlessly
• Market-data handler for full-depth order book and trade prints, normalised to my internal structures
• Order manager that submits, modifies and cancels in line with venue risk limits, returning acknowledgements and fills to the strategy in deterministic time

Acceptance criteria
• End-to-end latency benchmarked on local loopback and on the venue’s certification environment
• 100 % message-loss recovery verified through exchange replay or drop tests
• Clear build/run documentation so I can redeploy in a colocation rack without hand-holding

If you have hands-on experience wiring HFT engines to equities venues and squeezing every microsecond out of network stacks, I’d love to see your approach.