Why I joined Pylon: turning loan data into recursive self-improvement


I recently joined Pylon as a Member of Technical Staff, building the evals and data systems that make our AI more intelligent with every loan. People have asked me: why mortgage, and why Pylon? Four reasons: the work, the market, the team, and the moment.
The work
Pylon runs the whole loan, from application intake to capital markets settlement, so our data spans every customer segment, product, and stage of origination. Turning it into value is where I come in. I am building adaptive eval datasets from our historical loans to measure and improve the models and agents across the stack. Those evals power agentic development cycles that improve our pipelines continuously, freeing our people for the most critical and rewarding problems. Every loan that goes through Pylon improves the platform for the next one.
A massive market at the right time
Mortgage is one of the largest markets in the world, and one of the most fragmented. The gaps show up in how loans reach borrowers and in the 20+ systems lenders run on. This industry has not changed in decades, now it’s being completely rewritten.
I have a track record of picking traditional verticals and doubling down: utilities, SMB insurance, construction, and now mortgage. They are rarely the trendy choice, but each one sits on decades of data locked away in a fragmented tech stack and value chain. That is where AI can have the biggest impact.
I also look for companies with a truly unique approach: one platform that can reshape an entire industry. Pylon is building America’s mortgage rails: AI-native infrastructure that turns origination, from application through capital markets settlement, into a series of API calls.
People who execute
Vision only matters if the team can execute.
It starts with leadership. Leaders who have been there before know what to expect the second time around. I saw it at Next Insurance, and I see it again in Trent and Evan, who have both led and scaled teams. Evan and I immediately swapped war stories about scaling insurance companies through the same demanding regulatory environment.
Our leaders also get their hands dirty, which is critical when AI is changing how we work every few months. Josh, our CTO, still builds. When he found a bottleneck in how we ship, he rebuilt our code review and merge system himself. He also gives the team free rein to build its own tools, including a system that lets our AI agents communicate with each other.
The market has responded
In my conversation with Trent, it was clear that Pylon had reached a tipping point. Adoption is accelerating, and demand keeps growing. That is a great problem to have.
Mortgage has run on paperwork and handoffs for decades. We are building the rails that let it run on data. If you are an engineer or researcher who wants to help build them, we are hiring, and I would love to hear from you.