- Location
- San Francisco, California, United States
- Job type
- Permanent
- Specialism
- Other Software
- Employer
- Confidential
- Posted
- October 8, 2026
Product Engineer, AI-native insurance | SF Bay Area, 5 days onsite | $180K–$280K + bonus + equity
Commercial insurance still runs on email, PDFs and phone calls. Our client is a brokerage that rebuilt all of it on its own AI stack, from prospecting to quoting to binding.
Series A, around 60 people, growing very fast, with an engineering team of 18.
This isn't a ticket queue. You own business numbers. You sit with sales, service and underwriting, find the expensive problem, and ship an AI fix, often the same week.
What you'll do
- Own one or more business KPIs and move them with AI-powered features
- Build agentic systems (RAG, MCPs, LLM orchestration) across the full stack
- Prototype, ship and iterate in hours, not weeks
- Support what you've built in production, including on-call
What they're looking for
- 4+ years of full-stack engineering shipping AI-powered production systems (7+ for the staff level)
- AI agents or LLM features in production with real users. Not demos, not POCs
- Daily use of AI coding tools (Cursor, Claude Code or similar)
- Front end, back end and the AI layer, end to end. AWS or GCP
- Python and TypeScript preferred, but they're open on languages
- You can explain a technical tradeoff to someone who isn't technical
Bonus points for Series A–C experience or a fast promotion record, and for voice AI or real-time systems.
Probably not for you if your work has been infra or backend only, or you prefer a slower, process-heavy pace.
The detail
- Onsite in the SF Bay Area, 5 days a week, with an early start
- $180K–$280K base + performance bonus + equity. Senior or staff is decided in the interviews
- Process: two rounds. A 60-minute technical (a deep dive on something you've built, plus system design), then a superday. Come ready to talk tradeoffs and failure modes
If you've shipped an AI feature people actually use, send me a message. What was it, and how many people use it?
Commercial insurance still runs on email, PDFs and phone calls. Our client is a brokerage that rebuilt all of it on its own AI stack, from prospecting to quoting to binding.
Series A, around 60 people, growing very fast, with an engineering team of 18.
This isn't a ticket queue. You own business numbers. You sit with sales, service and underwriting, find the expensive problem, and ship an AI fix, often the same week.
What you'll do
- Own one or more business KPIs and move them with AI-powered features
- Build agentic systems (RAG, MCPs, LLM orchestration) across the full stack
- Prototype, ship and iterate in hours, not weeks
- Support what you've built in production, including on-call
What they're looking for
- 4+ years of full-stack engineering shipping AI-powered production systems (7+ for the staff level)
- AI agents or LLM features in production with real users. Not demos, not POCs
- Daily use of AI coding tools (Cursor, Claude Code or similar)
- Front end, back end and the AI layer, end to end. AWS or GCP
- Python and TypeScript preferred, but they're open on languages
- You can explain a technical tradeoff to someone who isn't technical
Bonus points for Series A–C experience or a fast promotion record, and for voice AI or real-time systems.
Probably not for you if your work has been infra or backend only, or you prefer a slower, process-heavy pace.
The detail
- Onsite in the SF Bay Area, 5 days a week, with an early start
- $180K–$280K base + performance bonus + equity. Senior or staff is decided in the interviews
- Process: two rounds. A 60-minute technical (a deep dive on something you've built, plus system design), then a superday. Come ready to talk tradeoffs and failure modes
If you've shipped an AI feature people actually use, send me a message. What was it, and how many people use it?
Make your move.
Send your CV and we’ll take it from there. No hard sell. No spam.
