AI & Automation Leadership · Biopharma R&D
Senior AI and automation leader in biopharmaceutical R&D, and a data-science PhD. I turn ambitious AI into systems that pass audit, earn trust, and actually ship.
What I do
Designing AI that does real work end to end — with humans kept firmly in the loop where it counts.
Turning GxP, 21 CFR Part 11, and emerging AI regulation into controls teams can actually operate.
Combining automation, ML, and LLMs to remove manual effort without removing accountability.
Setting the architecture, economics, and governance that let AI move past the pilot and last.
Agentic AI is an escalation pattern, not a default. Exhaust the simpler approach first — reach for the complex one only when the value genuinely requires it.
— Operating principle
Portfolio
Interactive tools and explainers on agentic AI decision-making, unit economics, and AI-first architecture — all live on GitHub Pages, one click to open.
“Should we build this?” — a six-step decision walk-through for agentic AI initiatives. Answers the questions cost models don't ask, and returns a structured view of whether to build, configure, buy, or defer — and which governance bodies belong in the loop first.
An interactive companion for understanding agentic AI cost levers: six business-friendly sliders, three preset scenarios, three pricing surfaces. Moves the conversation from “AI is expensive” to “here are the levers.”
A demonstration of disciplined AI architecture decision-making: choose the simplest pattern that solves the problem, and price the true cost of an agentic system before building it.
An interactive, client-side explainer that turns a static target-architecture slide into an explorable experience — documenting the future state of an AI-first life-sciences platform and making the semantic AI layer (ontologies → knowledge representation → models → inference) explicit.
Live Applications
Full-stack applications deployed on Render, built for regulated-industry workflows — open one and use it right now. Not screenshots: the real thing.
A production-grade LLM evaluation and human-in-the-loop review application for regulated industries — GxP-aligned quality monitoring across pharma R&D use cases, packaged for one-click deployment on Render.
An AI-powered proof of concept that analyzes clinical-trial Informed Consent Forms to classify participant consent for future and secondary data use. Returns structured, audit-ready results with confidence scoring and human-review flags — decision support, not legal advice.
Upload a cancer metabolite panel as CSV — the app auto-resolves each compound to SMILES via PubChem, then predicts co-folded 3D protein–ligand structures with the NVIDIA NIM OpenFold3 API.
A production-grade Flask application for pharmaceutical supplier onboarding — replacing a manual, nine-system procurement workflow with an intelligent orchestration layer powered by Claude.
Thought Leadership
The latest writing, pulled automatically from my Substack — where AI, automation, RPA, and BPM meet biopharma, from agentic systems and governance to M&A and valuations.
Domain
I work where AI has to be both ambitious and accountable — pharmaceutical R&D, pharmacovigilance, and other settings governed by GxP and 21 CFR Part 11. That means evidence over hype, controls over shortcuts, and architectures that hold up when the context gets long and the audit is real.
Focus — GxP · 21 CFR Part 11 · Pharmacovigilance · AI governance · NIST AI RMF · ISO 42001 · LLM evaluation & HITL · FinOps for AI · CoE design
Connect
Open to conversations on AI strategy, governance, and agentic systems in regulated industries — and the occasional good problem. Pick a slot below, read my writing, or reach out on LinkedIn.