I build, operate, and invest
in AI companies
Product leader turned entrepreneur with a knack for translating deep tech into business outcomes. I started as an ML researcher on autonomous vehicles at Ford, then moved to the business strategy side — because I realized the hardest problem in AI isn't the model, it's the go-to-market.
As Director of Product at Untether AI (later acquired by AMD), I led product strategy for next-generation AI accelerator hardware. After my MBA at Harvard Business School, I founded Guddi Growth, where I evaluated 200+ companies — including 50+ AI startups — developing pattern recognition for what separates AI companies that scale from those that don't.
I think about AI through three lenses: Can you build it? Can you sell it? Can it compound?
DealScope AI
AI-powered platform for analyzing Confidential Information Memorandums. Persistent deal pipelines, automated financial modeling, sensitivity analysis, and professional exports. Built for PE/VC analysts and search fund operators.
AI Startup Investing
Selectively investing in early-stage AI companies where I can add value on product-market fit, GTM strategy, and the bridge between technical capability and commercial traction.
Agent Economics
What does an AI agent actually cost to run? An interactive calculator for agent unit economics: cost per task, per-user COGS, and margin curves across Claude, OpenAI, and Gemini list prices. Companion to The Inference Tax.
Run the numbers →I write about what I see at the intersection of AI, product, and investing — the patterns, the pitfalls, and the counterintuitive truths that emerge when you evaluate hundreds of companies. All my long-form writing lives on Substack.
The Utilization Assumption
The largest compute contracts in tech history rest on a number set in sprint planning, not at signing. Why utilization — shaped by product and engineering choices around context, batching, and routing — matters more than any price negotiation.
Read on Substack →The Inference Tax
The hidden cost layer that fundamentally constrains what AI products can do. Why inference literacy — knowing how to route across model tiers rather than defaulting to frontier — is becoming the most important skill in AI product engineering.
Read on Substack →The Utilization Gap
Despite $600 billion in AI infrastructure spending, GPU clusters average just 5% utilization. Why memory walls and network bottlenecks — not raw compute — determine inference economics and which AI businesses are actually profitable.
Read on Substack →The Barrel Upgrade
How AI is reshaping workforce hierarchies — not by replacing workers, but by upgrading lower-tier talent into high-output contributors. A new lens for evaluating AI's impact on teams and organizations.
Read on Substack →The Concentration Thesis
Why AI inference capabilities will concentrate among a handful of dominant players rather than commoditize across competitors. The structural dynamics that separate long-term AI winners from the rest.
Read on Substack →The $16 Trillion Flip
AI services revenue looks enormous on the surface — but the unit economics are built on sand. An analysis of the fragility hidden beneath the current AI adoption wave and what it means for investors.
Read on Substack →The Pruning Principle
Why the best companies deliberately kill initiatives that are working. On strategic elimination as a growth discipline and the counterintuitive art of saying no to good opportunities.
Read on Substack →When Distribution Isn't Enough
Massive distribution used to be an insurmountable moat. AI is breaking that assumption. Why established platforms with billions of users are losing to startups with none — and what it means for GTM strategy.
Read on Substack →The Conviction Tax
Championing contrarian ideas before the market catches up comes with a hidden personal and professional cost. On what it really means to be early and right in investing and product strategy.
Read on Substack →Building in Public, Judging in Private
The best products are built in private and revealed fully formed. Why the build-in-public movement optimizes for attention over excellence, and what great product builders actually do instead.
Read on Substack →The Taste Gap
In a world where execution is becoming cheap, discernment is the new competitive advantage. On taste as a product skill, why judgment matters more than ever, and how to develop it deliberately.
Read on Substack →Let's talk
Whether it's about AI product strategy, a startup you're building, an investment opportunity, or just a good conversation — I'm always up for it.