About

I'm an AI engineer. I build advanced AI systems for clients across VC, media, education, finance, and energy.

I'm especially interested in the intersection of epistemology and AI — how we can build systems that don't just give answers, but show their work and surface the strongest counter-arguments.

Currently: Head of AI to Alex Epstein, and I run TW Consulting.

What I build

I build advanced AI systems that work in production, and I help teams take existing AIs that “almost work” across the line. I take full ownership of every project, and I have a small team of engineers who support larger AI systems in production.

I'm Head of AI to Alex Epstein, where I build AlexAI, a public AI energy and climate expert grounded in Alex's own work; AlexAI Pro, a more powerful, private version for internal users and policy makers; and a few other internal products for Alex and his team.

I also run external audits of AI and machine-learning systems, reviewing architecture, data pipelines, and evaluation instrumentation.

If you have an AI system you want built or audited, see my work and get in touch to see if we're the right fit.

Background

I have a degree in computer science and have been building AI systems since 2020 — knowledge bases, research pipelines, agents, and the evaluation tooling that makes them reliable. Before that, I founded and sold a software and media startup, EventIgnite, which gave me an appreciation for building things people actually want to use.

What I Think About

Truth-Seeking AI

Verification is what makes the 10x-100x productivity gains in software engineering possible — agents can keep running, check their own work, and iterate. I'm building the equivalent for intellectual work: AI that stress-tests arguments, surfaces counter-arguments, and provides assurance that claims have been examined.

AI Capabilities and Limits

AI optimizes within a given context. It finds better tactics, sharper evidence, more effective phrasing. But it doesn't step back and ask: "Am I arguing the wrong thing entirely?" Humans break out of context. AI executes within it. Understanding this distinction is key to using AI well.

The 100x Public Intellectual

AI can now do most of what used to require a research team. The bottleneck shifts to clarity — knowing what questions to ask, what arguments to stress-test, what's worth pursuing. The people (and AIs) who provide that clarity will have outsized impact.

Connect

I'm always interested in talking to people working on similar problems. Reach out on Substack, X, or LinkedIn.