I build to unlock capability. Start with a need, then I plan out what it takes to create a resolution. I work to keep things simple without compromising the standard. My projects span real-time data, trading, and simulation systems, built full-stack using agentic coding workflows.
Stack: Python · JavaScript/TypeScript · Node.js · FastAPI · WebSockets · PostgreSQL · MongoDB · Godot/GDScript · Linux/Ubuntu · pandas/parquet · real-time market data (Databento, CME futures)
- Trade Chart — Real-time futures replay simulator + live monitor. A FastAPI/WebSocket sidecar streams 1-second OHLCV for 15 CME contracts, with volume-splice continuous-contract construction, rolling percentile-rank metrics, and threshold alerting. Python · JS
- EDA Tool — Client-side exploratory analysis for trading event data. Walk-forward validation with immutable version history and an auto-optimizer that enforces strict in-sample / out-of-sample separation. JavaScript
- Armed Existence — Deterministic, config-driven RTS in Godot 4. Fixed-timestep seeded simulation (multiplayer-ready), ~350 JSON-tunable parameters, clean sim/render separation. GDScript
Selected private work: real-time speech-recognition trading bot (live broadcast keyword detection → automated event-market orders); Linux node-operation automation with health-check and auto-restart bots.
