Portfolio
Production-facing industry work, independently implemented LLM systems, ongoing private engineering, and peer-reviewed graph ML. Project maturity and evidence are labeled clearly.
Sole AI engineer: problem formulation, pose and segmentation workflows, anatomical measurement, validation, and product-facing API integration.
Document ingestion, indexing, retrieval, inference routing, structured traces, automated evaluation, and focused tests.
Extending an LLM/RAG foundation into an application-focused system. FastAPI, Docker Compose, n8n, agentic AI, MCP, and CI/CD belong to the planned scope; feature completion is not publicly verifiable.
A mathematically motivated graph-coarsening framework developed and evaluated collaboratively for molecular property prediction.
Related experience
My broader work includes knowledge-graph embeddings and link prediction with PyKEEN, predictive modeling for healthcare, and graduate instruction. I have also studied football video analytics concepts, without claiming a deployed football analytics product.
Research background → · Professional history →