Technical Program Manager, Supercomputing

Thinking Machines Lab Inc.; San Francisco (CA); On-site; USD 350,000 - 475,000; Full time

Benefits offered by this job

Health, dental, and vision coverage

Unlimited PTO

Paid parental leave

Relocation support

Job summary

Thinking Machines Lab Inc. in San Francisco invites you to drive compute allocation for frontier AI research. You will own allocation decisions, work with researchers and infrastructure engineers, and translate priorities into practical plans under constraints.

You will teach teams to plan compute needs, improve scheduling and utilization, and develop tools to make recurring decisions. This role rewards clear reasoning, collaboration, and the ability to adapt as workloads evolve.

Qualifications

• Experience making resource allocation decisions in a complex computing environment with clear ownership of tradeoffs and outcomes.
• Strong understanding of distributed systems, ML infrastructure, or HPC; ability to reason about workload interactions and constraints.
• Ability to explain tradeoffs and constraints with clarity and provide practical guidance.
• Ability to teach complex technical ideas through clear explanations and concrete examples.
• Experience mentoring researchers, engineers, or TPMs in compute planning and resource management.

Responsibilities

• Own compute allocation; understand demand, develop allocation plans, and carry decisions through to execution.
• Optimize within constraints across capacity, hardware suitability, workload size, timing, and dependencies.
• Improve productive use of compute; collaborate to close gaps between allocated resources and useful work.
• Teach teams to estimate needs, understand options, and reason through tradeoffs with real workloads.
• Develop examples, guidance, and tools to enable recurring decisions.
• Connect research and infrastructure; align plans with usable capacity and future needs.
• Learn from outcomes; revisit assumptions as workloads and priorities evolve.

Skills

Resource allocation

Distributed systems

Machine learning infra

Teach complex ideas

Data-driven decisions

Cross-team collaboration

Attention to detail

Tools

GPU clusters

Scheduling systems

HPC infrastructure

Job description

The mission of Thinking Machines is to build AI that extends human will and judgment. We are training frontier models with Inkling, developing Tinker to let people make models their own, and crafting interfaces that broaden human-AI communication. We believe the future worth building is human, and we're hiring people who want to build it.

About the Role

Compute allocation is a technical judgment problem. Research teams have different goals, workloads have different requirements, and available resources rarely match every need at once. Progress depends on understanding those constraints and finding ways to accomplish more within them.

In this role, you'll own compute allocation and help us make better use of our supercomputing resources. You'll work closely with researchers and infrastructure engineers to understand demand, evaluate competing needs, and translate research priorities into practical allocation decisions. You'll investigate where constraints limit progress and help teams find workable alternatives.

Teaching is central to this work. You'll help researchers plan their compute needs, understand allocation decisions, and reason through tradeoffs themselves. By working through real workloads, explaining your decisions, and developing useful tools and guidance, you'll make specialized knowledge available to more of the team.

This role is a good fit if you enjoy working deeply with technical constraints, making decisions under uncertainty, and helping others develop their judgment.

What You'll Do

• Own compute allocation. Understand current and upcoming demand, develop allocation plans with technical leads, and carry decisions through to execution.
• Optimize within constraints. Work through tradeoffs across capacity, hardware suitability, workload size, timing, and dependencies. Help teams adapt their plans as requirements and resources change.
• Improve productive use of compute. Partner with researchers and engineers to investigate the gap between allocated resources and useful work. Identify opportunities to improve scheduling, workload placement, and utilization.
• Teach teams to plan and use compute well. Help researchers estimate their needs, understand available options, and recognize the consequences of different choices. Build practical understanding through real workloads and decisions.
• Make your reasoning transferable. Explain assumptions and tradeoffs clearly. Develop examples, guidance, and tools that enable others to handle recurring decisions and recognize when they need help.
• Connect research and infrastructure. Keep research plans grounded in usable capacity and make future needs visible to supercomputing and infrastructure teams.
• Learn from outcomes. Compare expected and actual usage, revisit assumptions, and improve allocation decisions as workloads, systems, and priorities change.

Skills and Qualifications

• Experience making consequential resource allocation or capacity decisions in a complex computing environment, with clear ownership of the tradeoffs and outcomes.
• Strong technical understanding of distributed systems, machine learning infrastructure, or high-performance computing. You can reason about how workload requirements interact with system constraints.
• Demonstrated ability to improve useful output from limited resources. You can explain what constrained the work, which alternatives you considered, and why your approach helped.
• Ability to teach complex technical ideas through clear explanations and practical examples. You have helped others become more capable and independent.
• Comfort investigating usage and performance data directly, questioning assumptions, and changing your view when the evidence warrants it.
• Sound judgment when teams have competing priorities. You earn trust, explain difficult decisions, and work constructively through disagreement.
• A hands-on, collaborative approach with strong attention to detail and follow-through.

Preferred qualifications:

• Experience owning compute allocation at an AI research lab or another organization with substantial shared computing resources.
• Familiarity with GPU clusters, distributed training, scheduling, and workload placement.
• Understanding of how networking, storage, reliability, and hardware differences affect usable capacity.
• Experience building tools or analyses for capacity forecasting, allocation, or utilization.
• Experience mentoring researchers, engineers, or technical program managers in compute planning and resource management.

Logistics

• Compensation: The expected annual salary range for this position is $350,000 - $475,000 USD.
• Visa sponsorship: We support visa sponsorship and work with candidates through the process.
• Benefits: Health, dental, and vision coverage; unlimited PTO; paid parental leave; and relocation support as needed.