Product Lead for IBM's Agentic AI Platform. I work on the enterprise agent control plane: orchestration, evaluation, governance, and the shared context layer that lets teams run AI agents with confidence rather than optimism.
The short version of how I got here: Bronx kid. Started out climbing telephone poles for AT&T. Then healthcare compliance software, where being wrong gets counted in patients, not points. Then HashiCorp, building reliability and change-risk systems for a nine-figure cloud portfolio, until IBM bought the company partway through. New badge, same laptop 😂.
The part of the work nobody photographs. The layer that decides what runs, what stops, and who answers for it. I read the incident reviews before I read the strategy deck.
A few things I keep coming back to:
- Start where it breaks. Incidents and workarounds tell you what the roadmap missed.
- Measure before you promise. If you can't tell whether the system improved, it's a demo, not a product.
- Constraints are the brief. Regulation, latency, cost, headcount. Design around them honestly.
- Build for the person on shift. The operator at 2am is the real user. Everything else is a preference.
- construct — outcome engine. State an outcome, a learning staff of roles fills in the rest.
- blackstory — pins Black history to the places it happened. Archiving what would otherwise go unrecorded.
- geralddagher.com — personal site and writing. Next.js, more atmosphere than a personal site strictly needs.
Husband, father, South Florida. Foodie as a whole personality trait. Powered by coffee and an amount of Red Bull I'm not going to put a number on 😂.