Lead Product Manager, Agent Platform

Description

The company is building a vertically integrated mortgage platform that turns loan origination into a single API. Most other financial products have already been reduced to an API call, but mortgages are still assembled by hand.

The US mortgage market is worth $13T and underwrites the middle class. Public companies and some of the largest originators already build on the platform, and revenue grows every month.

Investors include Peter Thiel, Conversion Capital, QED, Citi, Fifth Wall, and the founders of Ramp, Blend, and Mercury.

The team is largely former founders who move fast, debate openly, and own results without needing direction.

Agents already handle much of the mortgage work, both inside the company and at customers. This role owns the Agent Platform, the base every internal agent runs on and the route outside agents use to work with the company.

Internally, you will build the tools and infrastructure that engineers, operators, and business teams use to build, test, ship, and monitor agents for document reading, underwriting, and closing. Externally, you will make it the easiest place for customers' agents to do real work.

Responsibilities

• Decide where agents will have the biggest impact, and align product, engineering, and operations on that.
• Set the roadmap for the platform's core parts: tools, orchestration, context and memory, retrieval, evals, observability, and guardrails.
• Cut the time internal teams need to take an agent from idea to production, and provide shared patterns so no one rebuilds the same thing.
• Run evals as a product in their own right, with shared datasets, benchmarks, and regression checks, plus human review wherever a regulated decision needs it.
• Shape how outside agents connect, including MCP servers, APIs, scoped permissions, rate limits, and documentation that both people and LLMs can follow.
• Ensure every agent action can be traced and explained to regulators, designed in from the start.
• Make build or buy decisions on models, vector databases, frameworks, and vendors, watching cost, latency, and quality.
• Write and ship production code, build agents on the platform, and use it the way its users do.
• Work with GTM and Account Management to connect customers' agents, support deals, and make sure customers succeed.
• Help hire and grow the team as the platform scales.

Requirements

• You hold clear opinions, defend them with reasoning, and change your mind when shown you are wrong.
• You rank priorities ruthlessly, keep the team on the highest-return work, and turn down good ideas that come at the wrong time.
• You are at ease in a heavily regulated market, where compliance is not negotiable and rules are followed while aiming to win.
• You care deeply about customers and borrowers, and you will do what it takes for them to succeed.
• You prefer software to manual work and treat any human step in backend mortgage processing as a bug. When a temporary workaround is needed, you log it and push for the proper fix.
• You build real relationships with internal teams, customer engineers, and many model and tooling vendors to get results.
• You generate ideas about the product and the team and act on them.
• You dig into details. You learn mortgage well and follow AI research closely, because you build with it.
• You have 4+ years as a product manager.
• A plus: an engineering background.
• You have owned and scaled a developer platform, infrastructure product, or API used by several teams or outside customers.
• You have shipped LLM-based agents or agent tooling to production and understand tool use, MCP, multi-agent orchestration, and structured outputs.
• You know evals, including golden datasets, LLM-as-judge, and regression testing, and you can tie model quality to business results.
• You understand retrieval: RAG, embeddings, vector databases, chunking, hybrid search, and reranking.
• You know the tradeoffs among prompt and context engineering, model choice, and fine-tuning, and how each affects cost, latency, and accuracy.
• You know tracing, guardrails, prompt injection, and how to protect sensitive data such as borrower PII.
• You write code, can commit to a production codebase, build prototypes, and go deep on architecture with engineers.
• You have built 0-1 enterprise software in a regulated industry.
• You can write SQL and define metrics that show whether agents work and whether teams use the platform.
• Nice to have: experience with mortgage or fintech infrastructure, or document understanding work such as OCR, extraction, and classification.
• You do not need to be a mortgage expert on day 1, but over time you must become one of the most trusted mortgage experts on the team to do this job well.

Conditions

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