I design and ship agentic workflows, LLM-powered apps, and serious data pipelines. Some of my deepest work is in research settings at GW; the same stack shows up in products, eval harnesses, and OSS.
Selected impact
Who I am, in a nutshell.
I work at the intersection of AI engineering and data science: agents and retrieval systems, evaluation and modeling, and the product discipline to get ideas into something people can run. I’m at GW’s Center for Global Mental Health Equity (School of Medicine), with earlier roles as a grad RA in engineering and a stint on DOE-linked analytics — so I’m used to both research-grade rigor and shipping pressure.
Publicly you’ll see LLM safety tooling, RAG stacks, planning/education products, benchmarking in ARES, and plenty of glue code. Domains shift; the through-line is reliable AI systems and honest data work.
Portfolio last updated: 2026-03-24
AI engineering and data science in research and product contexts, plus earlier industry and internship roles.
AI Engineer Data Scientist LLM & RAG Agents & eval
Ship end-to-end AI workflows for large research programs: retrieval and agent-style pipelines, smaller models and LLM integrations, evaluation harnesses, and production-minded glue so systems survive real studies — not just demos.
Owned modeling and analytics for multi-partner programs (including WHO/UNICEF-linked initiatives): metrics and study design support, ML and reporting pipelines, data quality, and turning messy operational data into decisions stakeholders trust.
Dashboards and analytics for DOE-funded vehicle and EV adoption research (the kind of work that ended up on VehicleTrends-style public views).
Built data science and small software pieces for sustainability analytics so teams could report faster and argue from cleaner numbers.
Default mix spans agents, products, and OSS. Second tab deep-dives healthcare & regulated-AI repos.
Open source and public artifacts
Each blurb has links out to GitHub or LinkedIn so you can poke at repos, PRs, commits, or posts instead of taking my word for it.
What I reach for when I’m building something real, not a slide deck.
Agentic workflows, tool use, RAG, prompt/eval loops, benchmarking (e.g. ARES), and shipping LLM features without mystery behavior in prod.
Modeling, measurement, experimentation, classical ML, and the storytelling that turns results into decisions.
Python, FastAPI, Django, pipelines, SQL, APIs, async when it earns its keep.
React, Next.js, TypeScript, Vite, UX when an AI is in the loop.
AWS, Azure, Docker, CI/CD, GitHub Actions, monitoring so deploys aren’t scary.
Formal training that backs up the messy self-taught parts.
Aug 2024 to May 2026
Nov 2020 to Jun 2024
Always happy to chat about collaborations, research, or a product role if it’s a fit.