High-Frequency Stock Momentum Engine
Budget: ₹250,000 – ₹500,000 INR
I want to commission a purpose-built high-frequency trading engine focused exclusively on stocks and driven by a momentum strategy. The system must ingest real-time streaming market data, react in microseconds, and execute orders with minimal slippage.
Key expectations
• End-to-end architecture: ultra-low-latency market data capture, signal generation based on momentum indicators, smart order routing, and risk controls.
• Exchange connectivity: direct API or FIX gateways to major equity venues; the code has to be exchange-agnostic enough to let me plug in additional venues later.
• Deterministic performance: consistent round-trip times under heavy load, with benchmarks and profiling evidence.
• Safety nets: position limits, kill-switch, and detailed logging that can be toggled without adding latency.
Deliverables
1. Source code with build instructions (C++ or another suitable low-latency language).
2. Configurable momentum model (look-back windows, thresholds, and adaptive sizing).
3. Automated tests and a replay utility so I can validate behaviour against historical tick data.
4. Deployment guide for a colocated Linux server, covering kernel tweaks and NIC settings.
Acceptance criteria
• Average end-to-end latency from tick arrival to order entry ≤ 50 µs in a lab test.
• No memory leaks or GC pauses under a 24-hour stress run.
• Successful execution of a provided test plan that includes volatile market scenarios.
If you have an existing framework you can adapt, let me know what parts we’d reuse; otherwise, be ready to build from scratch and document every component clearly so I can extend it later.
Key expectations
• End-to-end architecture: ultra-low-latency market data capture, signal generation based on momentum indicators, smart order routing, and risk controls.
• Exchange connectivity: direct API or FIX gateways to major equity venues; the code has to be exchange-agnostic enough to let me plug in additional venues later.
• Deterministic performance: consistent round-trip times under heavy load, with benchmarks and profiling evidence.
• Safety nets: position limits, kill-switch, and detailed logging that can be toggled without adding latency.
Deliverables
1. Source code with build instructions (C++ or another suitable low-latency language).
2. Configurable momentum model (look-back windows, thresholds, and adaptive sizing).
3. Automated tests and a replay utility so I can validate behaviour against historical tick data.
4. Deployment guide for a colocated Linux server, covering kernel tweaks and NIC settings.
Acceptance criteria
• Average end-to-end latency from tick arrival to order entry ≤ 50 µs in a lab test.
• No memory leaks or GC pauses under a 24-hour stress run.
• Successful execution of a provided test plan that includes volatile market scenarios.
If you have an existing framework you can adapt, let me know what parts we’d reuse; otherwise, be ready to build from scratch and document every component clearly so I can extend it later.