william-burks.dev · LinkedIn · will@william-burks.dev
Software engineer at JPMorgan Chase, Markets Technology (NA Rates). Day job is trade management systems in Python and React; before that, two years of Java/Spring and Kotlin backend in consumer banking. Outside of work I build market-data infrastructure, AI agents, and MCP servers, and I'm learning low-latency C++.
QuantWorkstation —
end-to-end algo trading workbench. ArcticDB bars store, vectorbt research
harness with IS/OOS walk-forward gating, prop-firm-style risk engine
(daily loss halt, trailing drawdown, symbol and total exposure caps), and a
scheduler daemon. Crypto via Alpaca, futures via IBKR, paper by default.
Includes 8 role-specific Claude Code agents under .claude/agents/ —
opus/sonnet/haiku tiered by effort, MCP-integrated, orchestrated.
finbert-sentiment-trader — 2024 sentiment-trading experiment. Archived as a post-mortem on position sizing and signal-quality failures; the risk engine in QuantWorkstation is the direct response.
markowitz-api — FastAPI service for Markowitz mean-variance portfolio optimization. Solves for max-Sharpe allocation via SciPy SLSQP across 10k random portfolios on the efficient frontier.
unified-search-mcp — MCP server that unifies semantic search across papers, experiments, hypotheses, and notes in my research workspace. FastMCP + ChromaDB + SQLite FTS5.
- NeetCode 150 on the C++ track
- Reading Meyers (Effective Modern C++), Williams (C++ Concurrency), Ghosh (Algorithmic Trading Systems)
BS Molecular Genetics + CS minor, Ohio State (2024). SWE I → SWE II at JPMorgan in 18 months.


