Principal Product Manager - Inference Engine

Role overview A principal product leadership role shaping the strategy for a GPU-powered inference business serving developers and AI-native companies. The position owns GPU allocation, pricing, packaging, and utilization, while partnering deeply with engineering to set the roadmap for serverless, dedicated, and batch inference offerings. It is suited to a product leader with strong technical judgment who can balance developer experience with infrastructure economics at scale.

Responsibilities - Own GPU strategy for the inference business, defining how capacity is deployed, allocated, priced, and optimized across serverless, dedicated, batch, and future offerings. - Build a product and business framework that improves token revenue per GPU hour, reduces idle capacity, and lifts gross margin as the business scales. - Define the inference product roadmap in partnership with engineering, covering prompt caching, autoscaling, batching, latency optimization, observability, dedicated deployments, compliance, and media model support. - Balance developer experience with infrastructure economics, making rigorous tradeoffs around latency, availability, throughput, pricing, and cost-to-serve. - Create pricing and packaging strategy with finance, go-to-market, and engineering, including SKUs, discounting, and packaging for serverless, dedicated, and enterprise customers. - Work directly with AI-native startups, mid-market customers, and strategic accounts to understand model needs, performance, compliance, and deployment patterns. - Partner with engineering and infrastructure to translate customer demand into GPU fleet planning, model serving, capacity allocation, and reliability improvements. - Establish operating metrics such as GPU utilization, token throughput, revenue per GPU hour, latency, error rates, model adoption, margin, retention, and capacity efficiency.

Requirements - Deep product judgment in infrastructure or AI, with experience building infrastructure, developer platforms, ML platforms, inference systems, cloud services, or other highly technical developer products. - Strong understanding of GPU economics, including utilization, throughput, latency, CapEx, cost-to-serve, gross margin, capacity planning, and workload placement. - Fluency in modern AI workloads such as LLM inference, open-source models, model serving, prompt caching, batching, model routing, media models, and production AI patterns. - Technical depth paired with business orientation, able to work credibly with infrastructure engineers while making clear tradeoffs for executives and go-to-market teams. - Strong analytical rigor, comfortable building frameworks and models that turn ambiguous signals into clear product direction. - Customer-obsessed mindset focused on developers and AI-native companies, with a recognition that infrastructure products must be reliable, performant, simple, and economically sustainable. - Executive-level communication and an ownership mindset suited to fast-moving, ambiguous environments where the product category is still forming.

Benefits and work setup - Hybrid role with compensation range of $218,000–$273,000 plus potential bonus and equity, including equity grants upon hire and an Employee Stock Purchase Program option. - Reimbursement for relevant conferences, training, and education, plus access to a large library of on-demand learning courses. - Employee Assistance Program, local meetups, and a flexible time off policy, with specific benefits varying by location. - Equal opportunity employer with a global, growth-oriented engineering culture.