Technical Product Manager - Soperator
Core
Own product direction for Soperator, a Slurm-on-Kubernetes control plane for GPU clusters, enabling ML engineers to run, scale, and optimize distributed workloads.
Role type
Technical Product Manager (AI Infrastructure)
Builds
Soperator control plane for GPU clusters
Domain
AI Cloud Infrastructure / Distributed Systems
Deliverable
product features
Required skills
Product Management, ML infrastructure/MLOps, distributed systems, cloud platform engineering, Slurm, Kubernetes, Ray, product analytics, data-informed prioritization
Preferred skills
GPU platforms/HPC primitives (InfiniBand/RDMA, topology-aware scheduling), modern ML training stacks (PyTorch, DeepSpeed, FSDP/ZeRO, NCCL), efficiency/reliability metrics (Goodput, MFU), observability, SRE/reliability engineering, customer-facing technical experience
Technologies
Slurm, Kubernetes, Ray, InfiniBand, RDMA, PyTorch, DeepSpeed, FSDP, ZeRO, NCCL
Responsibilities
Own the full user journey across Soperator clusters including workflows, dashboards, alerts, node lifecycle, and capacity management; Define product direction end-to-end from problem discovery to adoption; Lead deep customer discovery through interviews and usage analytics; Drive execution across platform teams (compute, networking, storage, observability, IAM); Translate frontier ML and infrastructure ideas into practical product capabilities; Define success metrics and prioritize roadmap decisions; Lead the open-source strategy and execution for Soperator