Lead Product Manager, Agent Platform

The posting, as Pylon wrote it

About Pylon

America’s $13T mortgage market is one of the most important financial systems in the world. It underwrites the middle class and is the mechanism through which millions of families build wealth. But while every other financial instrument has been simplified to an API call, mortgages are still assembled by hand.

We started from zero and created the first vertically integrated mortgage platform that turns origination into a single API.

Publicly traded companies and the country’s largest originators are already building on Pylon. Revenue is compounding monthly. We’re backed by Peter Thiel, Conversion Capital, QED, Citi, Fifth Wall, and the founders of Ramp, Blend, and Mercury.

Working at Pylon isn’t for those seeking comfort. The people who thrive here have high agency, strong opinions, and a track record of delivering outcomes without direction. Many of us are former founders. We move quickly, challenge each other directly, and take full ownership of results. It’s hard work, but it will be worth it.

Join us in building America’s mortgage rails.

ABOUT THE ROLE

Agents are starting to do a lot of the work in mortgage, both at Pylon and at our customers. We’re hiring a Lead PM to own the Agent Platform, which is what every Pylon agent is built on and how outside agents work with Pylon.

The job has two halves. The first is internal. You’ll build the tools and infrastructure our engineers, operators, and business teams use to build, test, ship, and monitor agents, for everything from reading documents to underwriting to closing.

The second is external. Our customers are building their own agents, and you’ll make Pylon the easiest place for those agents to do real work. That means MCP servers, APIs designed with agents in mind, and the permissions and guardrails that keep it all safe.

What you decide here will shape how quickly we can automate the mortgage process, and how easily the rest of the industry can build on us.

WHAT YOU’LL DO

• Set the direction for agents across Pylon and get product, engineering, and operations aligned on where they’ll make the biggest difference
• Own the roadmap for the platform’s core pieces, like tools, orchestration, context and memory, retrieval, evals, observability, and guardrails
• Make it much faster for internal teams to go from an idea to an agent running in production, and give them shared patterns so nobody builds the same thing twice
• Treat evals as a product of their own. Teams should have shared datasets, benchmarks, and regression checks, plus human review wherever a regulated decision needs it
• Own how external agents work with Pylon, including MCP servers, APIs, scoped permissions, rate limits, and docs that both people and LLMs can follow
• Make sure we can trace and explain every action an agent takes. We’d rather design for regulators from the start than bolt it on later
• Make build or buy calls on models, vector databases, frameworks, and vendors, and keep a

[…]
ow how to build real relationships and get results through them.

• You have a lot of ideas. You notice things that could be better about the product and the team, and we’ll give you room to act on them.
• You’re willing to get into the weeds. You can learn mortgage well enough to sound like a 20-year veteran, and you keep up with AI research like someone who actually builds with it.

REQUIREMENTS

• 4+ years as a product manager. An engineering background is a plus
• You’ve owned and scaled a developer platform, infrastructure product, or API used by several teams or outside customers
• You’ve shipped LLM-based agents or agent tooling to production and know your way around tool use, MCP, multi-agent orchestration, and structured outputs
• You know evals well, including golden datasets, evals, LLM-as-judge, and regression testing, and you can connect model quality to business results
• You understand retrieval, including RAG, embeddings, vector databases, chunking, hybrid search, and reranking
• You understand the tradeoffs between prompt and context engineering, model choice, and fine-tuning, and how each one affects cost, latency, and accuracy
• You’re familiar with tracing, guardrails, prompt injection, and protecting sensitive data like borrower PII
• You write code and are comfortable committing to a production codebase, building prototypes, and going deep on architecture with engineers
• You’ve built 0-1 enterprise software in a regulated industry
• You can write SQL and define the metrics that show whether agents are working and whether teams are using the platform
• Nice to have: mortgage or fintech infrastructure experience, or work on document understanding (OCR, extraction, classification)

We don’t expect you to be a mortgage expert on day 1. Over time, though, you’ll need to become one of the most trusted mortgage experts on the team to do this job well.

Must live in or be willing to relocate to Palo Alto (CA) or New York City.