Point of View
Knowledge is commoditized.
The new currency is judgment, agency, and decision power.
Most people use AI for the mundane—fixing typos, catching bugs, reviewing PRs.
The ones who are really leveraging it work differently. They spend more time on architecture and judgment, then let AI execute at scale.
The human provides conviction. The machine provides velocity.
This is happening everywhere, not just in code. The question is no longer can AI execute the work? It's can you trust it?
"In code, you can verify.
In finance, you need conviction.The Thesis
Here's the thing: one error destroys trust in everything. If page one has a mistake, you have to validate the entire document. At that point, you might as well have built it yourself. The time savings disappear.
The companies that win will be the ones building trust layers for AI execution—systems where you don't have to blindly believe, where every assumption is traceable, where errors surface instead of hide.
What I'm Building
Precurion
Trust layer for AI execution in finance.
Financial analysts can model at AI speed, but build with conviction. Every input is sourced to the document. Every assumption is traceable. You don't blindly trust—you verify.
When a model says the asking price is $5M, where did that number come from? What page? What contract? Is that contract even real?
We're building verification layers that go deeper than "here's a source." Is it contextually correct? Surface errors early, make assumptions traceable, so you never have to review everything from scratch.
Excel was built 40 years ago for humans. Its positional encoding—A1, B2—is optimized for human eyes, not AI.
Language models don't see grids. They see tokens.
Background
Y Combinator
S23 Batch
Cursor
Worked w/ current COO
Tinkerer
Loves the work
Non-linear path to building AI products
Started in sales—built and led teams of 10-12 people. That sparked an interest in tech, which led to a pivot into data science.
Spent years in the trenches: marketing attribution models, churn prediction, likelihood-to-convert systems. Built custom analytics dashboards for enterprise clients. Learned what actually matters when you're shipping to real users with real stakes.
Moved into fintech, then AI-powered content production, then enterprise intelligence. Each step closer to where AI creates the most leverage—and where trust matters most.
Sales
Leadership
Data
Science
Fintech
YC S23
AI
Products
Precurion
Now