At the University of Toronto, I build AI-enabled scientific software that connects LLM applications, machine learning services, experiment orchestration, data pipelines, and web interfaces. I also build CI/CD and evaluation workflows to investigate failures across software services, model outputs, networking, and laboratory hardware.
I built HELIOS, a multi-agent system connecting scientific reasoning and experiment planning with workflow generation, validation, execution, monitoring, and recovery. Its benchmark evaluation separates decision quality from execution reliability, with typed contracts, validation gates, and traceable decisions.
My PhD research at Western University focused on biomedical machine learning, including real-time surgical modeling, optimization, and MRI enhancement. I also develop retrieval-based AI applications, MCP tools, and Bayesian optimization services, and bring technical documentation and code review into my work.
