Danilo Campos

Robopsychologist

I conceive and build LLM tools operating at scale. I teach the adventurous to seize a new frontier.

The Wizard

Architected and built PostHog's LLM agent-powered integration wizard. Thousands of delighted users onboarded each month. Architecture extensible enough to support a whole team's contributions.

The Accords

Creator of the Agent Accords—shared understanding frameworks for human-AI collaboration, built on the Agent Skills specification. Guidance that serves both humans and AI agents from the same document.

Got a problem nobody can name? Let's talk.

mail@danilocampos.com

"I just ran the posthog install wizard and I am blown away by how amazing it was. It genuinely feels like magic."

— Matt White, YC founder

"Hot take: the best AI dev tool I've used this month is... a setup wizard."

— Ben Warren, Mesa.dev
PostHog Wizard ad on a San Francisco bus stop
San Francisco, February 2026

The ground has moved

The cost of machine intelligence is collapsing. Every software product will be reorganized around agents that can read, reason, and act. There are no playbooks for any of this yet. I can help you write yours—from the perspective of building AI agents in production.

Approach

Every project in my career follows the same pattern: read the culture, build the intervention that belongs inside it, watch it stick. The medium changes—mobile apps, hardware, AI agents—but the method is the same. I make technology culturally compatible.

What I do

Advisory & Workshops

I teach select teams to build AI agents that work—not from theory, but from the practice of having shipped one at scale. The obstacle is rarely technical. It's cultural: teams that built their identity around one paradigm struggle to adopt the next, not because they lack skill, but because the new tools don't fit who they are yet. I make them fit.

My goal is to make your team self-sufficient, not dependent on me. I genuinely love this work—I've been building social-technical bridges since I was a kid with a borrowed Macintosh.

Speaking

Upcoming: Monkigras (London, March 2026), AI Engineer Europe (April 2026). Topics include AI agent architecture, MCP design, and the paradigm shift in software engineering. Available for conferences, corporate events, and team sessions

The Agent Accords

Open-source frameworks (CC BY-SA) that give humans and AI agents shared vocabulary and principles. Install in Claude Code or read on the web.

Notable work

Writing