Engineering reliable systems around machine learning

I build reliable AI systems and backend infrastructure for complex technical workflows.

At Oak Ridge National Laboratory, I engineered asynchronous FastAPI services that orchestrated multi-agent reasoning, retrieval, and generative-model workflows. At Blueprint, I build production software and data-driven routing logic for real operational users. My work at Carnegie Mellon focuses on modular LLM experimentation and automated evaluation.

My EECS and Bioengineering background also informs large-scale bioinformatics pipelines at UCSF and scientific machine-learning research at the Garcia Center. Across those settings, I focus on service boundaries, reproducible pipelines, failure handling, and systems that remain dependable beyond a model demo.