Hi, I’m Marc!
I’ve spent most of my career building software products and the teams that build them—moving between individual contributor and leadership roles across startups, consultancies, and mission-driven work. These days, I focus on applying generative AI in product: helping companies figure out where AI can genuinely improve what they’re building, then doing the work to make it happen—whether that’s advising executives, building systems myself, or educating engineers.
Recent Articles
View all →-
Microagent Architecture
Monolithic LLM agents break down under the weight of bloated contexts, tool confusion, and prompt sprawl. Microagent architecture borrows the discipline of microservices: explicit contracts, per-agent evals, decomposition under pressure, while reckoning with new physics: lossy handoffs, probabilistic components, and a meter that never stops running.
-
Do we still need TDD?
Agentic coders can’t practice traditional TDD—but the underlying values of rapid feedback, testability, and scope discipline matter more than ever. Forcing red-green-refactor onto agents is performative; preserving the underlying values requires rethinking the process and rituals.