Quant Developer — Automated Trading System (Interactive Brokers, live-trading experience required)

Quant Developer — Automated Trading System (Interactive Brokers, live-trading experience required) Overview: I'm building an automated trading system to trade ETFs, futures, and options through my own funded Interactive Brokers account. I'm looking for an experienced quant developer to build and validate it. This is an engineering-and-validation role, not a "bring me a winning strategy" role — I supply the strategy logic; you build the infrastructure, implement the strategies, and validate them honestly. The project starts as a single strategy on one instrument, taken all the way through validation and paper trading, before expanding to all three instruments and to regime-based strategy switching. For the right person there's a clear Phase 2 (futures, options, regime switching) and a longer runway. Required skills Interactive Brokers API — hands-on. Not "I've used broker APIs" — actual IBKR: the TWS API / IB Gateway (or the Client Portal Web API), placing and confirming orders programmatically, and handling IBKR's specific quirks (keeping the Gateway alive, the periodic re-authentication, and reconnecting/re-syncing state after a drop). Python. The primary language for this build, with IBKR libraries like ib_insync or the native ibapi. Resilient long-running services, not scripts. You can build a system that runs 24/5 without silently dying: proper error handling, reconnection logic, logging, and state that survives a crash. Risk-system design. Kill switch, position sizing, daily loss limits, and circuit breakers — including knowing what happens to an open position when the connection drops. This is the code that protects the account. Honest strategy validation. Out-of-sample / walk-forward testing, realistic transaction-cost and slippage modeling, and a genuine understanding of overfitting. Reconciliation logic. Continuously confirming the system's view of positions matches IBKR's. Instrument-specific knowledge: Futures: contract rollover, tick/point values, margin. Options: the Greeks, options-chain handling, wide-spread / liquidity-aware order placement, assignment and expiration. ETFs: including leveraged-ETF decay awareness. Regime detection / regime-based strategy selection — and understanding that the selector itself must be validated, not just the individual strategies. Backtesting frameworks — custom, or experience with QuantConnect / Backtrader / Zipline. Monitoring & alerting — heartbeats and push notifications so a failure doesn't go unnoticed. Nice-to-have skills Testing discipline — automated tests for trading logic (unit + integration), especially around order handling and the risk module. Idempotent order handling — preventing duplicate or double-fired orders from retries or reconnects. Backtest-to-live parity awareness — understanding why backtest results diverge from live (slippage, fill assumptions, look-ahead bias) and how to minimize the gap. Observability — structured logging and monitoring/alerting experience. Data engineering — clean handling and storage of market data (time-series DBs, dealing with gaps and bad ticks). Config-driven design — strategy parameters and risk limits in config files rather than hardcoded, so they can be adjusted without a code change. Version control fluency (Git). Basic VPS / deployment / DevOps — process managers, auto-restart, secrets management. Statistics depth beyond backtesting, supporting validation rigor. What I'm not looking for HFT / low-latency / FPGA work, market making, statistical arbitrage, crypto/DeFi, and institutional portfolio construction. This is a retail automated-trading system on IBKR — none of that applies. If your expertise is centered there, it's likely not a fit, and I'm not looking to pay institutional rates to over-engineer a straightforward system. Machine learning as a core approach. This system is rule-based by design — that's intentional, and often more robust here. No ML premium needed. The single most important thing: have you built a live IBKR system before, and can you tell me specifically what broke and how you fixed it? That matters more than any list above. What you'll build: Integration with the Interactive Brokers API (TWS/IB Gateway or Client Portal Web API), deployed to run reliably on a VPS. A risk module and automated kill switch (daily loss limit, per-trade sizing, max positions) — built first, before profit-seeking logic. Order management with reconciliation (system state always matches the broker), plus logging, monitoring, and alerting. Honest backtesting on out-of-sample data with realistic costs, for both individual strategies and a regime-based strategy selector. A paper-trading run on IBKR before any live deployment. Full documentation and an operational runbook (how to start, stop, recover, and monitor the system) — required for final acceptance. You must have: Demonstrable experience building live automated trading systems (not just backtests). Direct Interactive Brokers API experience strongly preferred; equivalent broker-API experience considered only if you can speak specifically to how you'd handle IBKR's Gateway uptime, re-authentication, and reconnection/state-resync behavior. A rigorous approach to out-of-sample validation and a real understanding of overfitting. Experience with risk controls and failure handling — kill switches, reconciliation, reconnection after dropped connections. Experience handling futures (contract rollover, tick/point values) and options (Greeks, options-chain data, wide-spread order handling) — or a clear, credible plan for how you'll approach them. Secure credential handling — trade-only API keys (never withdrawal-enabled), stored in a secrets manager, never hardcoded or shared in plaintext. Terms: Fixed-price, milestone-based (design → validated backtest → built system → passing paper-trading run → live rollout support). Payment on delivery and acceptance of each milestone. All code and IP assigned to me, with an NDA. Non-negotiable. Go-live is gated on paper-trading results meeting pre-agreed criteria — not on a fixed date. Please state your weekly availability and expected timeline, and confirm you'll do the work personally (no subcontracting without disclosure). To apply, answer these five questions (applications without answers will be ignored): Describe a live trading system you built. What broker/API, and what did it trade? Walk me through exactly how you validate a strategy to avoid overfitting. Tell me about a time a system you built failed for a technical reason (dropped connection, bad fill, bug). What happened and what did you change? How would you design a kill switch that reliably closes positions and locks the system on a daily loss limit? This will start as a single strategy on one instrument before expanding to all three and to regime switching. How would you phase the build?