CV

A concise view of my experience, projects, education, and technical focus.

AI engineer · Applied machine learning

I build and evaluate AI systems from training loop to production workflow.

My work spans LLM post-training, inference engines, coding agents, enterprise retrieval, and evaluation systems for clinical and internal knowledge workflows.

3+ years Software engineering and ML
71 SAPs Clinical document generation
4 systems Post-training, agents, inference, vision

Experience

Applied AI in production settings

Atrium · Pfizer engagement

AI Engineer

  • Built a clinical document-generation system that retrieved protocol evidence from Neo4j and produced structured statistical analysis plan sections.
  • Added Pydantic contracts, OpenAI Structured Outputs, and an LLM-as-a-judge evaluation layer to catch missing evidence, contradictions, and unsupported additions.
  • Reduced first-draft generation from one to two days to 1.0-3.4 minutes across 71 plans and contributed to the associated peer-reviewed study in Clinical Trials.

Tezo

Software Developer (ML)

  • Built enterprise data pipelines in Snowflake with controlled promotion across development, staging, and production environments.
  • Engineered behavioral and temporal fraud features and trained an XGBoost classifier that reached 78% fraud recall at 42% precision on held-out data.
  • Deployed batch inference through Amazon SageMaker and connected fraud labels and risk scores to Snowflake and Power BI for analyst review.

Selected projects

Systems built to understand the full stack

Education

University of Maryland

MS, Applied Machine Learning
College Park · 2026

Deep learning, NLP, advanced ML, reinforcement learning, optimization, probability, and statistics.

Technical focus

BuildPython, PyTorch, Triton, SQL, FastAPI, Pydantic

TrainFSDP, LoRA/QLoRA, Transformers, DeepSpeed, W&B

ShipDocker, Kubernetes, SageMaker, Azure ML, Snowflake, CI/CD