About the role
structured by ORIBack to all jobs Product Manager - AI Hybrid Published 1 month ago About Us Phyllo is a data gateway that allows social data to be accessed from source platforms (e.g. YouTube, Instagram, TikTok, Twitch, Upwork, Shopify, and more).
What you will do
- Own the roadmap for Clayface: which business questions we answer well next, which types of analysis we take on, and what we deliberately leave out.
- Design how multi-step workflows fit together, which steps the system decides for itself, and where a person should stay in the loop.
- Set the bar for what a good answer looks like and work with engineering on how it gets checked.
- Make sure every number can be traced back to a source and a time period.
- Partner closely with engineering, design, data, sales, and customer success.
What they are looking for
- 5 years in product management, including at least one AI product or feature that real users used and that you measured after launch.
- You are hands-on with AI products. You have written and improved prompts yourself, looked closely at what the system produced, and worked out why it went wrong.
- A working understanding of how AI products are put together today: how a model gets the context it needs, how it uses tools and data sources, how multi-step workflows are strung together, and where these systems usually break.
- Real experience with evaluating AI output. You have set up ways to test quality systematically and used the results to decide what to change.
- You care about evidence. You check whether an answer is genuinely supported by the data, and you say so when it is not.
- Comfortable with data. You can write SQL or work in a BI tool, and you are appropriately suspicious of numbers that look too clean.
- Enough technical depth to have a real conversation with engineers about trade-offs, cost, speed, and data dependencies without needing everything explained from first principles.
- Clear writing. A good spec from you should prompt specific questions about edge cases rather than confusion about what you meant.
- Comfortable saying no, including to people who are excited about something that demos well.
Nice to have
- A background as an engineer, analyst, or data scientist
- Experience with consumer goods, retail, or market research data: syndicated data, retailer sell-through, digital shelf, or social listening.
- Experience building products that sit on top of messy data from many different sources.
- Familiarity with tools for testing and monitoring AI products.
- Experience with B2B or enterprise customers, including what their security and procurement teams ask for.
- An advanced degree in a technical or business field.
Before you apply
- If you have worked on an AI product before, tell us about something it kept getting wrong and what you did about it.
Benefits
- Work from home or your preferred location
- Flexible hours
Full posting text
Experience: Associate