One person. Twenty years.
A working conviction that pharma is where AI has the most to prove — and the most to gain.
I’ve spent two decades inside regulated pharma — systems management, quality, and product ownership— the parts of the industry where “it has to be right” isn’t a slogan, it’s the job.
Now I build AI tooling to the same standard, in the open. epiCorpus is where that work lives — and where I make the case that regulated industry is uniquely equipped to lead on AI, not fear it.
The record.
What I’m building.
Each opens a detailed case study ↗
PRJ-01epiClavis — compliant AI engineThe flagship: an engine reaching toward GAMP / clinical-software requirements, using AI to drive consistency.FlagshipPRJ-02Spanish-language learning appA commercial product, built and shipped end to end.CommercialSVC-01epiCursor — web studio for local businessMy web-design service for small local businesses.In production ↗The philosophy.
Pharma treats AI as a risk to contain. I think that’s backwards — the discipline pharma already has is exactly what makes AI safe to adopt.
Proprietary software, built better.
AI can raise the speed, cut the cost, and hold the quality of the bespoke software biotech runs on — without lowering the bar.
The work is pigeonholed.
Working in Claude Code, tasks are scoped and auditable — the opposite of the data free-for-all people imagine when they hear “AI.”
Reaching the clinical-software bar.
Projects like epiClavis push AI toward the regulatory standards clinical software is held to — and use it to drive consistency.
Fewer licences, more bespoke.
Point AI at the right places and in-house tools replace expensive vendors — and give staff services built for how they actually work.
Let’s talk.
The hiring funnel — and the vendor pitch — both have AI at every layer. This is the shortcut past it, straight to a conversation.