I build and test ML infrastructure, with a focus on inference systems, reliability, and reproducible evaluation.
I am a computer science major and mathematics minor at NYU, expecting to graduate in May 2027. I am interested in Fall 2027 PhD research in ML systems, as well as backend infrastructure and developer-tools engineering.
- Hugging Face Transformers #47238: fixed half-precision
torch.compilefailures in DETR-family position embeddings and added a compiled-dtype regression test. - marimo #10100: restored completion fallback after Jedi analysis failures and removed duplicate analysis.
- Home Assistant: improved Tuya startup recovery, LG ThinQ service error handling, and LinkPlay discovery.
- AgentCI Guard: an experimental TypeScript static analyzer for GitHub Actions workflows that run AI agents. It models workflow permissions and reachability and emits diagnostics with explicit analysis limits. Release v0.6.2 · Demo.
- TraceBench: workload and measurement infrastructure for studying ML serving behavior, with reproduction instructions and scoped experimental evidence in the repository.
- CloudTune: an early ML infrastructure project for supervised fine-tuning on customer-operated infrastructure, failure recovery, and portable signed execution receipts. The public research preview includes an illustrative receipt-verification demo; the application and source remain private.
How should an inference change be evaluated when throughput, latency, output requirements, and run-to-run variation all matter? I am interested in building reproducible experiments and failure cases that make those trade-offs explicit.
I am seeking supervised ML-systems research opportunities and useful open-source implementation work.