Product Manager, Inference Platform

THE ROLE

Product at Baseten is a nascent function. Our company today has a strong engineering culture, is heavily customer-obsessed, and moves fast. We're building the product function now, and you'd be one of the people who defines it. You'll work directly with our founders and with some of the best systems and AI engineers, and you'll set the standard for what product looks like here. PMs at Baseten don't sit above engineers. You earn ownership by being technical, finding the truth in front of customers, building great cross-functional relationships, and shipping great product experiences.

Once a model is deployed, keeping it fast, reliable, and economical at scale is where production inference is won or lost. You'll own the surface that makes that happen: how deployments autoscale, how traffic is routed, how the system fails over, and how workloads scale across clusters and regions. You'll own these as products end to end, both how they work under the hood and how customers configure and observe them, and you'll set the roadmap that infrastructure and product teams alike can build toward. This space is still evolving. Think cloud infrastructure in the mid-2000s. Your job is to make it 10x easier to reliably scale and serve AI models in production, and to set the market standard.

This role is a great fit if you love getting deep in the technical details, and if you'd rather do whatever it takes to ship for customers than stay in the strategy, UX, or docs layer. You're drawn to platforms, systems, GPUs, models, and scaling infrastructure more than to applied AI, and you find real satisfaction in the foundational, sometimes unglamorous work that makes AI systems reliable and scalable at the infrastructure level.

EXAMPLE INITIATIVES

Take a look at these blog posts written by members of our Product and Infrastructure teams:

• How we built Multi-cloud Capacity Management (MCM)

• How the Baseten Delivery Network (BDN) makes cold starts fast

• Baseten brings AI video to life on Nebius

RESPONSIBILITIES

• Own how workloads scale and where they land, from autoscaling to demand (up under load, down to zero when idle) to a single placement policy that expresses region, compliance regime, and capacity preference.

• Give compliance-bound workloads right-of-way on sensitive capacity.

• Make production inference reliable by default, so every request reaches a healthy replica and rolling deploys never drop traffic.

• Define region-aware routing, with multi-region and active-active failover as first-class policy, plus health-aware recovery from stuck or bad replicas.

• Build the release engine beneath safe rollouts: the traffic-shifting that powers canary, shadow, and A/B, along with warm-ups, drain, and probes.

• Push the cost and performance frontier for serving AI at scale across latency, throughput, uptime, and cost efficiency.

• Drive a measurable decline in MTTR through self-serve incident management.

• Set the roadmap that infrastructure and product teams build toward, and own each capability end to end, from backend behavior through the customer-facing configuration and observability surface.

REQUIREMENTS

• 8+ years in product management, including deep experience with infrastructure, distributed systems, or ML serving.

• Fluency reasoning about scaling, routing, failover, and the cost/performance frontier, at a level that earns the respect of staff engineers.

• Experience owning capabilities end to end, backend through UX, rather than a single slice.

• Track record driving cross-team roadmaps and the dependencies beneath them.

• Comfort defining a category that doesn't fully exist yet.

NICE TO HAVE

• Hands-on experience with GPU infrastructure.

• Experience with Kubernetes.

• Experience with serving frameworks like vLLM, TensorRT-LLM, or SGLang.

BENEFITS

• Competitive compensation, including meaningful equity

• Paid parental leave

• Fertility and family-building stipend through Carrot

• (U.S. only) Company-facilitated 401(k)