SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Lambda is an AI cloud infrastructure company serving tens of thousands of customers from AI researchers to enterprises and hyperscalers. The company's mission is to make compute as ubiquitous as electricity.
As a Staff Product Manager on the Foundational Infrastructure team, you will own a meaningful part of defining where customer data lives on Lambda and how it gets to GPUs. On the storage side, you'll cover where weights, checkpoints, and datasets sit, how customers ingest and keep data current, performance guarantees for model loading, and capacity metering and pricing. On networking, you'll cover data movement at the throughput training and inference demand, tenant isolation and VPC for enterprises, customer network connectivity, traffic balancing for serving endpoints, egress control, and bandwidth sharing.
You will work across both storage and networking domains rather than specialize in one, following customer needs. Much of this is feature parity with mature clouds that Lambda does not yet have. However, as a neocloud (not a hyperscaler), the job requires judgment about which capabilities matter and which carry unacceptable cost. AI workloads move data in patterns these primitives were not designed for, and identifying where that difference should change the product is the more interesting half of the work.
Great product managers at Lambda are defined by insight (looking at data to determine what it means for customers and business), influence (winning over engineers, designers, executives, and partners without relying on authority), and execution (delivering ideas to customers and iterating based on measurement).
Key responsibilities include: getting close to customers and support to identify real blockers on data and connectivity; ranking and sequencing work across storage and networking; producing precise requirements and acceptance criteria; deciding metering, pricing, and packaging for storage and data transfer; working through builds with engineering teams and making mid-flight tradeoff calls; setting launch criteria and ensuring operational readiness; defining success metrics before launch and measuring adoption and real workload performance; taking positions on contested tradeoffs and owning decisions; and partnering with other product managers and teams to land pieces as one product.
Requirements:
- 7+ years of product management experience, including time on storage or networking products at a hyperscaler, neocloud, or comparable cloud provider
- Shipped infrastructure products that external customers ran production workloads on, in object, file, or block storage, or in virtual networking, load balancing, or connectivity
- Real depth in one of the two domains (storage or networking) and enough range in the other to make decisions, plus appetite to close gaps quickly
- Understanding of how data movement and network behavior show up as performance for distributed training and large-scale inference
- Owned pricing, packaging, or metering for an infrastructure product
- Ability to discern which parts of a hyperscaler playbook transfer to an AI cloud and which are cost without benefit
- Can define iterative plans moving an organization from current state to desired outcome
- Can turn ambiguous customer, technical, and commercial inputs into product definition that multiple teams execute against
- Communicate plainly, write well, and make decisions easier for people without shared context
Nice to have:
- Experience with object storage and S3-compatible interfaces, or parallel and shared file systems at high throughput
- Experience with block storage, snapshots, or storage path behind instance boot and recovery
- Experience moving bulk data into a cloud, including resume, integrity verification, and customer-visible throughput
- Experience with virtual private cloud networking, security groups, or private connectivity to customer networks
- Experience with load balancing, controlled egress, or edge protection in front of production serving endpoints
- Experience with high performance fabrics for distributed training (InfiniBand, RDMA, GPUDirect)
