AI platform engineer with 20 years building large-scale distributed systems. Currently focused on the infrastructure other engineers use to ship AI agents: agent SDKs, MCP tool integration, model gateways, and the evaluation stack that tells you whether an agent actually works.
Selected projects
| Project | Stack | What it does |
|---|---|---|
| aarogya-atlas | TypeScript, Python, FHIR, LLM tool calling | Tool-calling loop over 12 typed tools with a critic pass over FHIR-aligned data; answers cite source facility ids. Cloud inference in the hosted demo; an optional on-device mode runs photo triage and embeddings on local models. Live demo |
| gcp-financial-data-platform | Go, Airflow, dbt, BigQuery, Terraform | Reference architecture for financial data infrastructure: event ingestion, batch transforms, governed access and cross-region backups, with the infrastructure as Terraform modules |
| geochemical-dating | XGBoost, LightGBM, scikit-learn | 2nd of 31 on Kaggle. Two-branch ensemble for 3-class geological age classification, private Macro-F1 0.96988, 0.00104 behind first, and generalised upward on the private split |
Competitions
ARC Prize 2026 (Paper Track) · AI Agent Security: Multi-Step Tool Attacks (OpenAI) · BirdCLEF+ 2026 · AI Mathematical Olympiad Progress Prize 3 · NFL Big Data Bowl 2026 · Gemma 4 Good Hackathon · Hack-Nation 2026 · Orbit Wars
Stack
Python, Go, SQL, TypeScript · agent SDKs, MCP, tool calling, evals and tracing · PyTorch, XGBoost, LightGBM · AWS and GCP · Kubernetes, Terraform, Airflow, dbt, BigQuery, Pub/Sub, PostgreSQL
MSc Artificial Intelligence · Dallas-Fort Worth, TX

