Research Engineer, Infrastructure, Numerics
Thinking Machines Lab · San Francisco
Frontier labonsite$400k–$500k
At a glance
Highlights
- Build core systems enabling stable training of trillion-parameter models.
- Improve numerical foundations of distributed training for large-scale LLMs.
- Co-design model architectures and training recipes with research teams.
What they offer
- Generous health, dental, and vision benefits
- Unlimited PTO
- Paid parental leave
- Relocation support as needed
Ideal candidate
A collaborative builder who thrives at the intersection of research and systems engineering and cares deeply about making large-scale training stable, fast, and reproducible. You bring a bias for action and love working across teams to ship impactful infrastructure.
Requirements
- Bachelor's degree or equivalent experience in computer science, electrical engineering, statistics, machine learning, physics, robotics, or similar
- Understanding of deep learning frameworks (e.g., PyTorch, JAX) and underlying system architectures
- Thrives in highly collaborative environment with cross-functional partners
- Bias for action with initiative to work across stacks and teams to ship
- Strong engineering skills with performant, maintainable code and debugging in floating-point numerics, low-precision arithmetic, and distributed systems
- Familiarity with distributed frameworks, FP8/INT8/MX formats, open-source infrastructure, and large-scale model training (preferred)
- visa sponsorship
- Visa sponsorship available; committed to working through the visa process together
- hiring process
- Evergreen role reviewed continuously as new opportunities open; avoid reapplying more than once every 6 months; may also apply to separate project-specific postings.
Lab profile
About Thinking Machines Lab
Builds frontier AI that extends human will and judgment through customizable models and interfaces.
What they build
- Inkling frontier models
- Tinker model customization
- Human-AI communication interfaces
Known for
- Frontier model training
- Post-training and reinforcement learning
- Evaluation infrastructure
- GPU supercomputing and reliability
Posted · Last seen