I work on enterprise AI: I run the demos, and I stay on for everything after them, where a product has to hold up against real data, real users and a procurement team.
Before this I co-founded Styx, an evaluation platform for AI agents. Off the clock I'm usually outside with a camera.
Enablement Engineer at Draup, building Etter.ai's skills intelligence module with the CEO and taking it to enterprise customers.
Built the flagship skills intelligence module from zero alongside the CEO; now live with enterprise customers and a differentiator in three RFPs. Own three accounts prospect-to-production, across healthcare, technology and FMCG.
Talent intelligence for Fortune 500 accounts, informing $2M+ in workforce decisions. Turned customer feedback into a 37% pipeline accuracy gain.
Angel-funded agentic AI evaluation platform. 129 discovery interviews, five design-partner pilots, eval framework for agents in CI/CD.
SAR-imagery analytics for defense and disaster-response clients.
#!/bin/sh # Things I will happily lose an evening to: topics="libre software, and who really owns the model weights orchestrating a string of coding agents until one of them finally does the boring part where the AI space actually goes once the demos stop landing what satellite imagery reveals about a landscape under stress" while read -r topic; do test -n "$topic" && echo "yes, let's talk about $topic" done <<EOF $topics EOF # Reply time: usually a day. Reply length: usually too long.
Held back from the page source so bots and scrapers don't pick them up.