Building is no longer the bottleneck. Distribution is. I build the AI systems that close that gap for B2B revenue teams, and I write about what the work teaches me at derr.ai.
- Maestro AI Revenue System™. A full agentic revenue system I architected end to end. Eval-gated agents implemented directly into a team's revenue processes and workflows, live on real engagements. The public demo shows one of its plays, turning buying signals (funding, hiring, website intent) into pipeline. The build story →
- Delivery loop. A closed-loop client delivery system. Weekly sync transcript in, tiered task queue out, agent execution behind trust gates, next meeting's deck drafted before I sit down. An agent pod that works like a team (consultant, PM, engineer) with human gates wherever the work carries risk.
- Templatiz. AI-powered content OS. Curates top-performing content, generates templated variations, schedules distribution. 200 beta users, 700+ on the waitlist.
- derr.ai. My corpus. Weekly deep dives on GTM engineering, agent evals, and production AI systems, published where both people and answer engines can cite them.
- Agent security beyond permission scoping. Ephemeral credentials, trust boundaries, and enforcement at the infrastructure layer. System prompts are advisory.
- Self-learning agent systems. Feedback loops that promote patterns from episodic memory into standing lessons, so the system gets smarter with every run.
- Eval frameworks for production agents. Binary pass/fail quality gates that agents clear before they earn autonomy. Evaluators built from observed failures.
Surfing HB and overlanding off Highway 395 in my Jeep 🏄♂️