PM, Compute Platform

PADO•Washington, DC

2d•Remote

About The Position

PADO AI Orchestration, Inc. is seeking a strategic and technical product manager to lead its compute platform. The role owns the full product lifecycle, from identifying market opportunities and defining product vision through launching features and iterating on them. The ideal candidate is a builder who can work in a fast-paced, ambiguous environment and combine commercial judgment with an understanding of interactions between software and hardware systems. The product manager will define the compute platform strategy and roadmap, improve GPU utilization and compute throughput, and turn emerging research and technical capabilities into products adopted by customers. The role works across engineering, UX/UI, marketing, sales, and data science, and partners with customers on discovery, deployment, delivery, and ROI. The position also tracks product outcomes and customer needs to guide future investment and product iterations.

Requirements

4–10+ years of product management experience, particularly with complex technical data products.
Strong understanding of data center infrastructure (including servers, UPS, and cooling), power systems (including three-phase power and redundancy), or AI/ML frameworks.
Solid understanding of HPC and AI cluster orchestration concepts, including cluster management, scheduling, and monitoring.
Experience with workload schedulers such as Slurm, Kubernetes scheduling (Kueue, Volcano, or custom schedulers), or similar; understanding of queueing, priorities, preemption, gang scheduling, and fair share.
Record of turning ambiguous research or emerging technical capabilities into adopted products through partnerships, integrations, or open source.
Technical fluency in agent architectures, observability, evaluation design, and automated experimentation techniques.
Ability to use data and experimentation, such as A/B testing, to drive product adoption and business decisions.
Exceptional written and verbal communication skills, including the ability to explain technical concepts to engineers and C-suite stakeholders.

Nice To Haves

Experience running or building for large GPU clusters (hundreds to thousands of GPUs) for training or inference.
Familiarity with NVIDIA's cluster stack (NCCL, DCGM, Base Command, Run:ai, DSX) or AMD equivalents (ROCm, RCCL).
Experience with GPU partitioning and sharing (MIG, time slicing, MPS) or inference-serving frameworks (vLLM, Triton, Dynamo).
Background in power-aware scheduling, workload power profiling, or data center power and cooling constraints.
Experience with cluster telemetry and observability (Prometheus, DCGM exporter, time-series data).
Exposure to neocloud or GPU-as-a-service business models, including reserved and on-demand capacity.
An engineering or computer science degree.

Responsibilities

Define and execute a long-term compute platform roadmap with executive and engineering support, designing for multiple data center and AI infrastructure segments.
Define capabilities to improve GPU utilization and goodput, including smarter queueing and preemption, bin packing, fractional and multi-tenant GPU sharing, and power-aware placement.
Turn emerging research and capabilities, including agent-driven optimization and workload placement, into adopted products through partnerships, integrations, or open source.
Partner with sales to lead discovery sessions and develop credible ROI and TCO arguments tied to GPU utilization, compute throughput, and operating expense savings.
Manage and measure deployments across on-premises and cloud environments, interfacing with DCIM, BMS, and workload scheduling layers.
Work closely with engineering, UX/UI, marketing, sales, and data science; participate in scrums and sprint planning and remove roadblocks to product delivery.
Lead voice-of-the-customer efforts and competitive analysis to identify gaps and investment opportunities; advocate for client needs internally.
Ensure products and features are delivered with high quality and speed; represent PADO to clients, deliver products to clients, and act as a forward-deployed PM or engineer as needed.
Track success metrics, including adoption, revenue, and task success, to inform future product iterations.