Chris Pezza

The short version

I’m a career marketer and former Director of Demand Generation with unusual technical and creative range. I studied computer systems and business, then spent fifteen years managing marketing systems at early-stage startups, twice as the first marketing hire. I was never a software engineer.

Today I own GTM systems and AI at Coalesce. That covers marketing’s stack end to end (tooling decisions, the AI budget, design, development, product management, DevOps) and a partner posture with sales and RevOps on the systems their teams run on.

How the work gets done

I decide what gets built and why, write those decisions down, and direct AI agents to build it. I’m judged by outcomes.

A few months ago I held that syntax was delegable and architecture was not. I have since moved that line. Agents now take a real part in design, and what stays with me is the decisions and the verification. Decisions Are the Interface Now makes that argument and tests it on one of my own projects.

In practice that means specs before code, decisions recorded before they get made twice, and checks an agent cannot quietly change. I don’t read every diff. I build the system that tells me whether the work held.

Three lanes

• AI Engineering: RAG systems with citation-first architecture, judgment models, Claude skills and plugins, agent workflows.
• Dev Tools: open-source starters, Claude Code plugins and an ops dashboard, mostly under the Clownware studio.
• Web & Brand: websites and content work for employers and clients, plus my own personal and studio sites.

Now

Running GTM systems and AI at Coalesce: owning marketing’s stack, partnering with sales and RevOps on theirs, and managing production of the Agentic Data Engineering and Agentic Data Migration practitioner’s guides with Coalesce’s subject-matter experts. On the side: Clownbot, my private stack for running open-weight models behind one policy gateway; Ops, the dashboard that ranks every repo and vendor account by what needs my next hour; the Astro and Go Performance Starters; and a private research vault that the writing here is sourced from.