Faros AI
Product Manager, AI Native Initiatives
$200K–$250K
Redwood City · Hybrid · Manager · Full-time · Posted May 19, 2026
- Experience
- 6+ years in product management; ideally at least 2+ years building AI/ML or LLM-powered products.
- Education
- CS degree or equivalent engineering background (preferred).
About the job
Own and develop Faros AI’s AI-native product initiatives, including agent collaboration workflows, evaluation capabilities, and connections to engineering context.
What you'll do
- Form product hypotheses, prototype and ship products, and test them with customers.
- Investigate agent performance and develop approaches to evaluating results, models, tools, and context.
- Drive a product area end to end, with outcomes measured by customer satisfaction, engagement, and successful deals.
What they ask for
- 6+ years in product management, ideally including 2+ years building AI/ML or LLM-powered products; shipped at least one product or major feature from a blank page to customers.
- Engineering experience or deeply technical product experience, including ability to read code and discuss API design, system architecture, and prompt engineering.
- Hands-on experience with LLM APIs, embeddings, vector stores, tool use, prompt design, and evaluation.
- Ability to experiment, prototype, write, ship, and analyze data and agent logs.
Nice to have
- Experience in developer tools, engineering productivity, dev infrastructure, coding assistants, or agent platforms; startup experience is strongly preferred.
- Experience shipping at a Series A or B startup, working on human-agent collaboration, or building production evaluation frameworks; CS degree or equivalent engineering background.
Pay
USD 200K – 250K (base salary not specified)
Tools
LLM APIs, OpenAI, Anthropic, embeddings, vector stores, API design, system architecture, prompt engineering
AI focus
Develop AI-native products for engineer-agent collaboration, agent evaluation, and engineering-context integration; experiment with models and tools and improve agent performance.
Model-written summary · the original posting is authoritative