Selected projects

Work

I build machine-learning systems from the model outward: data and training, evaluation and failure analysis, backend architecture, and deployment.

Clinical sequence modeling2024–Present

TEDDY

A transformer trained from scratch to forecast first-occurrence diagnoses from longitudinal clinical event histories.

150M+raw clinical events tokenized
  • PyTorch
  • Transformers
  • CUDA
  • Clinical sequences
Retrieval & ranking2023–Present

MentorMatcher

A production-oriented mentor search system where evaluation led from an unstable LLM judge to a fine-tuned cross-encoder reranker.

  • FastAPI
  • RoBERTa
  • PostgreSQL
  • FAISS
Open-source performance engineering2024

decontX Optimization

Profiling and redesigning a scientific Python bottleneck reduced a large single-cell workload from more than 12 hours to minutes.

~99%runtime reduction
  • Python
  • Sparse matrices
  • Concurrency
  • Single-cell RNA-seq
Computer vision2024–Present

Endoscopy AI

Semi-supervised image and video pipelines for lesion detection, with active learning and temporal propagation used to extend limited labels.

3K+endoscopy images
  • PyTorch
  • ConvNeXt
  • CBAM
  • Active learning