MincaAI — Remote Latam, Mexico preferred About MincaAI MincaAI is an AI-native insurance software company. HQ at Station F (Paris) x Mexico, operations across LatAm. We build production AI systems for enterprise insurance carriers across three pillars — underwriting, policy issuance, and claims. Real LLM pipelines. Real enterprise clients. We are small, we move fast, and we ship into environments where being wrong is expensive.

The Role

You own product end-to-end for one of our three pillars. Discovery, spec, delivery, adoption. From messy client reality to shipped feature that carriers actually use in production. This is a tech PM role, not a stakeholder-management role.

You read code, you understand distributed pipelines, you argue architecture with engineers on the merits. You know the difference between "the LLM is wrong," "the prompt is wrong," and "the extraction schema is wrong" — and you know why each one changes the roadmap differently. Insurance is not a UX layer.

Underwriting logic, policy issuance flows, billing reconciliation — the domain complexity is real, and the product decisions live inside it. You will not be handed a domain expert to translate for you. You are expected to learn the business faster than anyone expects, sit with brokers and underwriters, and come back with specs that are right the first time.

Enterprise client reality. We operate a SILO model — each carrier engagement has different infra, different processes, different constraints. Some clients move fast, others arrive with a 200-page compliance checklist.

You pivot between these modes without losing velocity. There is no template we deploy blindly.

The Bar We are hiring for one profile: someone who gets to the bottom of a new problem faster than the room expects, and then does the work.

- Speed of comprehension. Dropped into an unfamiliar underwriting workflow, you have the mental model in days, not months. You ask the two questions that matter, not the twenty that don't.
- Extreme ownership. No handoffs, no "waiting on engineering," no "the client didn't specify." If it's your pillar and it's broken, it's yours.
- High output. This is a demanding role at a startup pace. Roadmaps move weekly, clients escalate, and things ship. If you need a stable backlog and a two-week spec cycle, this isn't it.
- Intellectual honesty. You kill your own features when the data says so.



You say "I was wrong" fast and cheap. If you're the person your last team routed the hardest, least-defined problem to — we want to talk. What You'll Own - Product roadmap for your pillar: quarterly priorities, sequencing, trade-offs. You defend it and you revise it.
- Discovery: client interviews, shadowing ops teams, mapping the presente workflow — not the one on the org chart.
- Specs and PRDs: precise, testable, engineer-ready. Edge cases, data contracts, and a real definition of "done" for LLM-backed features.
- LLM product decisions: which tasks are LLM-shaped, which are deterministic, where to spend inference budget, when to invest in evals vs. ship.
- Delivery: sprint planning with tech leads, unblocking, protecting scope, flagging risk before it's a fire.
- Client-facing product work: demos, discovery calls, escalations, requirement negotiation. In Spanish and English.
- Metrics & adoption: define success per feature, instrument it, and drive follow-through when adoption lags.
- Cross-pillar coordination: underwriting, issuance, and billing share data models and services. You align on shared surfaces. Must-Have - 7+ years in product management, at least 3 in B2B / enterprise / SaaS.
- Technical depth: you read Python, understand REST/async APIs, and can reason about Postgres schemas and queue-based pipelines. You don't ship code, but no engineer has to translate for you.
- Shipped LLM-backed product in production — not a prototype, not a demo. You understand prompts, evals, latency/cost trade-offs, and how LLMs fail quietly.
- Track record of learning a hard domain fast and shipping into it. Be ready to prove it with a specific example.
- Enterprise discovery & delivery: you've turned non-technical business users' reality into shipped software.
- Specs engineers actually praise. Precise, complete, unambiguous.
- Bilingual English + Spanish, professional level. Client calls in Spanish, product reviews in English.
- Uses AI tools daily (Claude, Claude Code, Cursor, ChatGPT, or equivalent)



for discovery synthesis, spec drafting, research, prototyping. Not curiosity — workflow.
- Startup-fit: no PM ops layer to hide behind. You draft, you decide, you move. Nice-to-Have - Insurance domain knowledge: underwriting, policy admin, claims, billing, broker/carrier relationships. Any of it. Strong bonus if you've shipped into a carrier.
- Consulting background (MBB, Big 4, boutique strategy) — structured problem-solving, client posture, and comfort in messy enterprise environments is a real plus.
- Experience with document-heavy workflows: OCR, extraction, structured output from unstructured input.
- Founder, founding PM, or early PM (first three) at a startup. Our Stack (for context) - Apps: Python / FastAPI microservices, Vite + React + TypeScript
- Data: PostgreSQL + pgvector, Redis, S3
- AI: OpenAI + Anthropic APIs, in-house evals, RAG pipelines
- Infra: Kubernetes (self-managed + OpenShift), AWS, GitHub Actions
- Product tooling: Notion, Linear, Slack How We Work - Small focused changes. No new tech debt. No new product debt either.
- If a client request conflicts with the product vision, you push back and propose better.
- Written by default. Decisions live in Notion, not in someone's head.
- Ship, measure, iterate. No feature graduates without an adoption signal.
- Speed is a feature. So is being right. We expect both. Success - Your pillar has a clear, sequenced roadmap the team believes in.
- Features ship on time with specs engineers didn't need to rewrite.
- Client adoption is measurable and moving in the right direction.
- LLM features have evals, cost budgets, and known failure modes — not vibes.
- You've retired at least one thing that stopped earning its complexity.
- Six months in, you're the person the founders bring into the hardest client conversation. Logistics - Location: Remote, Mexico preferred (LatAm considered).
- Languages: Professional English + Spanish required. French a plus.
- Timezone: Overlap with LatAm working hours + some Paris overlap for HQ syncs.
- Start: ASAP.

How to Apply Send your CV + a short note (5 lines max) to [email protected] answering one question: Describe the hardest domain you had to learn from zero — how long it took you to be dangerous in it, and what you shipped because of it.

- Generic answers get filtered out. Be specific.

📌 Senior Product Manager — Tech (México)
🏢 MincaAI
📍 México