AI Product Manager — Agentic Systems
About the Role
We're an early-stage fintech startup building an agentic AI layer that transforms fragmented debt and financial data into intelligent, actionable recommendations for consumers and lenders. We're looking for a Product Manager who has shipped real AI systems in production — not just prototyped them — and who understands the full stack from data pipelines to user-facing intelligence.
This is a high-ownership, high-ambiguity role where you'll work directly with engineers and founders to shape the product roadmap. If you've spent more time in production incident reviews than in design sprints, you'll feel at home here.
What You'll Do
- Own the product roadmap for our agentic AI layer, including multi-step decision workflows, data reconciliation pipelines, and downstream automation
- Define and drive evals, reliability standards, and guardrails for LLM-powered systems operating on sensitive financial data
- Translate complex data fragmentation problems (messy debt data across servicers, institutions, and loan types) into structured, scalable product solutions
- Work closely with ML and data engineers on RAG pipelines, agent orchestration, tool/action design, and observability
- Build escalation logic and fallback systems that handle edge cases gracefully in regulated environments
- Partner cross-functionally with compliance, engineering, and business stakeholders to navigate the constraints of fintech infrastructure
- Define success metrics that go beyond engagement — latency, accuracy, coverage, decision quality
What We're Looking For
- Proven experience shipping AI/LLM-powered systems in production — you've dealt with hallucinations, evals, latency tradeoffs, and reliability at scale
- Strong technical fluency: you can hold a substantive conversation with engineers about RAG, agent frameworks, API design, orchestration, and observability
- Experience in fintech, healthtech, medtech, or another regulated domain — you understand compliance constraints, messy data ecosystems, and operational risk
- Deep familiarity with agentic workflows: multi-step reasoning, tool use, escalation logic, not just prompt design
- A data-first mindset — you know that AI output quality lives or dies on the pipelines underneath it
- High agency and a builder mentality — you don't wait for a roadmap to be handed to you
Ideal Backgrounds
Candidates from the following backgrounds tend to ramp quickly into this role:
- AI Systems PM or ML Platform PM
- Fintech Infrastructure or Data Platform PM with strong AI exposure
- Risk, compliance, or operational automation PM
- PM from regulated industries (healthtech, pharma tech, defense tech) who has worked on AI-driven decision systems
What Won't Work Here
If your AI experience is primarily consumer-facing chatbot features without ownership of production reliability, operational complexity, or data infrastructure — this role will be a stretch. We need someone who has been paged at 2am because a model degraded, not just someone who has written a PRD about an AI feature.
Why This Problem
- Debt data is fundamentally broken. Consumers hold loans, credit cards, student debt, and mortgages across dozens of institutions with no unified view. Our system gathers and reconciles that data, evaluates repayment options and product eligibility, recommends next actions, and triggers downstream workflows — all through an agentic layer that has to be accurate, fast, and trustworthy. The PM who owns this will have an outsized impact on how the system thinks, what it acts on, and how much consumers and lenders can trust it.