We’re looking for a Technical PM to own the agent runtime and the platform that powers our agentic products — and the non-functional requirements that make those agents accurate, fast, reliable, and cost-effective at scale.
You’ll own how well the platform runs, not the end-user features on top of it. You’ll define the runtime architecture, own the product evals that prove our agents stay accurate as we scale and upgrade models, and set the bar for latency, reliability, cost, and observability across our agent runtimes.
This is an engineering-facing role — but it’s a product role at heart. You translate what customers and the business actually need into the architecture that delivers it: product judgment expressed through technical decisions. You’ll spend your time in architecture discussions and eval design alongside ML and backend engineers, not in design reviews. When you work with go-to-market, it’s to turn runtime scale, stability, and reliability into a story the market trusts. And you’ll track how the frontier of agentic engineering is moving — keeping us ahead of it and being a visible voice on it inside and outside the company.
If you’ve never written a prompt chain, debugged a pipeline, reasoned about token cost, or designed an eval, this isn’t the right fit.