Execution infrastructure, order flow & backtesting.
Trading-specific engineering — the systems that keep strategies running: exchange connectivity, execution logic, market-microstructure tooling, and the community-facing access layer around a live Sierra Chart trading room.
Quant Researcher & Platform Engineer — Futures Research → Live Execution
Private Platform — Remote
2025 – Present
Built a systematic-futures research platform (~334k LOC Python, ~70 strategy families across six futures markets) whose product is unbiased backtests: a written no-lookahead contract enforced by CI, random-noise controls on every result, immutable fingerprinted caches.
Promotion path ends in bit-level parity: frozen JSON models re-implemented as C++ Sierra Chart studies, proven bar-for-bar and trade-for-trade (100.00% parity, 0.000000 R difference) before live orders.
Live order layer with a single submission choke point, append-only order journal, post-restart stale-level refusal guard, and CRC-checked feature-state snapshots.
Built a bot that watches X/Twitter accounts for signal tweets and places options orders through Interactive Brokers (ib_insync); paper mode by default.
Safety layer: per-trade and daily spend caps, position ceiling, trade cooldown, off-hours signals queued until the market opens, env/file kill switch, exponential backoff on API failures.
SQLite state so a restart never replays a tweet or forgets the day’s spend; structured JSON logs and Telegram alerts on every order and failure.
Standalone options screener: option chains → short-put metrics (delta, days to expiry, mid price), scored and ranked into SQLite or Google Sheets.
Multi-timeframe structure, strategy tables and real-time chart updates from Rithmic.
Optional bracket-order execution (entry/SL/TP) to Rithmic when strategy signals fire; researched ML models (XGBoost) for feature-based signal classification.