CS + Economics @ UT Austin '28. Right now I'm studying human-AI trust research at UT Austin: Trust Prism.
Previously AI/ML @ State Farm, SWE @ TRNDG, H-E-B, Zonda
Most of what I build is AI systems plus the tooling that proves they actually work — evaluators, constraint solvers, retrieval pipelines, and the backends underneath them.
| What it is | Stack | |
|---|---|---|
| Strength Atlas | Evidence search engine for strength training. Crawls and normalizes coaching claims, then serves them with full source provenance in sub-200ms. 175+ programs, 476 claims, 83 sources, 95% test coverage. | Python · FastAPI · Postgres · Supabase |
| predict-strength | From-scratch ML pipeline testing how predictable strength progression is from competition-only data, with leakage-safe modeling and baseline/model comparisons. | Python · scikit-learn · XGBoost · pytest |
| StormOps Console | Agentic risk-analysis platform that turns weather and grid context into operational directives for energy operators. | Python · RAG · multi-agent orchestration |
| Pinpoint | Real-time crowd-safety platform for large venues — live density heatmaps and sub-2s EMS alerts across three user roles. | React Native · Node/Express · MongoDB · WebSockets |
| neetcode-gpt | A GPT built from scratch — attention, tokenization, and training loop, no framework abstractions. | Python · PyTorch |
| Nudge | IoT pressure monitoring for wheelchair users. ESP32 firmware reads a 16×16 sensor matrix over BLE into a mobile app and Supabase backend. | ESP32 · BLE · React Native · Supabase |
| PeCo | RAG-powered strength coach. Certified capstone for Google's Gen AI Intensive. | Python · Gemini · ChromaDB |