This is a hands-on senior IC role with real ownership, reporting to the Director of Product. You will work day to day with our search engineering and research teams, with the evaluation team, and with customer success, who bring us quality feedback from real accounts every week. Our search engineers are deeply experienced in large-scale retrieval, and the product team is small, so you will argue your case directly with the people building it. That only works if you have built at this scale yourself.
Your responsibilities will include:
Set the quality bar and release criteria for search: what good means per use case, and what a change has to clear before it ships. Our evaluation team owns the measurement, and you work closely with them on what to measure and why.
Own the depth, latency, and cost tradeoffs across our search products, and make them explicit and defensible.
Shape the direction of our index: what it needs to cover, how fresh it needs to be, and what quality bar it has to meet.
Build a point of view on where search infrastructure is going as agents, rather than people, become the main consumer, and turn that into a roadmap.
Know how we compare to the alternatives customers evaluate, and drive the work that closes the gaps that matter.
Work with customer success and directly with customers to understand where our results fall short in real workflows, and turn that into quality priorities.
Partner with the data and vertical product owners on what our corpus needs to contain and how results should be shaped per industry.
5+ years of product management experience, with a meaningful portion spent on high-scale search or ranking systems: web search, feed ranking at a large social platform, or a large ad engine.
A background as an engineer, ML engineer, or researcher on a production system in one of those areas. This is a requirement rather than a preference. You will be making tradeoff calls alongside senior search engineers, and you need to have built something at that scale yourself.
Direct experience shipping a search or ranking change end to end, including how it was evaluated and what happened after it shipped.
A strong data foundation: you can design an experiment, read eval results critically, and tell a real quality improvement from noise.
Product judgment to match the technical depth: you can decide what to build and what to leave, and defend it to both engineers and customers.
Strong understanding of LLMs, AI agents, and how retrieval fits into agentic workflows.
Comfort with ambiguity, moving between strategy and execution without losing the thread.
The ability to influence and align cross-functional stakeholders without formal authority.
