Software engineer with a quantitative finance background. Specialty in finance and compute markets.
I've worked deeply in compute market research and ran a pre-seed company that dove deep into compute as a derivative from December 2025 to May 2026. Excellent in AI integration and automation of workflows/internal functions. Looking for a junior software engineering role at a startup. Dallas, TX. Willing to relocate.
Sixteen months shipping software in prod, all of it in finance or compute. I entered the software market as a quantitative researcher at a Dallas fund, then dove into the world of startups with AI-centric CRMs for small businesses but once I saw a much bigger fish I knew I had to try and chase it: compute derivatives. I spent five months as the sole engineer and founder of a startup in that space, and now building the internal platform for a private equity fund as well as my own internal quantitative research tooling platform.
What I know best: quantitative strategy creation, financial index design and creation, data plumbing, normalizing heterogeneous accelerators into one unit. The compute work, from the unit to the September index rebuild, is written up in full on the compute page.
Sole engineer on the internal platform for a search-fund private equity practice. FastAPI backend in Python, TypeScript/React frontend. Ingestion for searcher applications and deal flow, with outbound and inbound delivery in production. Docker on self-hosted bare metal, continuous deployment through GitHub Actions runners, permission-gated data access for AI agents.
Designed the Compute Unit, a normalization metric across GPU, TPU, LPU and NPU accelerators. Provisional patent filed January 9, 2026. Built and ran the price pipeline behind the index: 54M+ price ticks across 28 providers at peak (15 reporting today) and 56 accelerator models across GPU, TPU, LPU, NPU and ASIC, on TimescaleDB and Docker, still collecting nine months later. Took product direction from Goldman Sachs futures sales. Raised from March, wound down in May after Silicon Data–CME and Ornn–ICE.
Researched and backtested strategies in a live fund environment. Built an RRG and PCA-based regime system over 63 years of daily prices and 13,674 tickers, and helped extend the firm's backtesting infrastructure. Under NDA.
Chain-linked geometric price index for GPU compute over listed rate-card prices, normalized to the Compute Unit. Replaced an index that averaged whatever reported that day. 54M+ ticks, 28 providers at peak (15 reporting today), 42 GPUs.
Measures what an accelerator delivers instead of trusting the spec sheet, across four dimensions. Locks clocks and power, waits for thermal steady state, trimmed mean with a median fallback under throttling. NVIDIA and AMD.
Daily price store for about 51,000 US-listed instruments with a security master keyed by FIGI. DuckDB over per-security parquet, Postgres for definitions. Deployed by a self-hosted GitHub runner on push. Extending the platform's functionality daily.
Relative Rotation Graph engine over 63 years of daily prices and 13,674 tickers. RS ratio and RS momentum for every ticker in the GICS hierarchy against PCA benchmarks, with drill-down from market to stock. Dash app and CLI.
AI voice-agent platform for small businesses. Answers inbound calls, qualifies leads, books appointments, emails the caller a confirmation with the Zoom link, texts the owner, and syncs Google Calendar. Built and operated solo. Call the agent at (617) 313-8175.
