Latest AI JobsBrowse Jobs
Thinking Machines logo

Research Engineer, Infrastructure, Kernels

Thinking MachinesSan FranciscoResearch Infrastructure (ML Infrastructure and Training Stack)
Apply on company site
Salary
$350K - $475K USD
Job type
Full-time
Work mode
Remote
Posted
4 Aug 2026
Status
Open · checked 7d ago

About this role

Thinking Machines is hiring a full-time Research Engineer, Infrastructure, Kernels in San Francisco with remote flexibility to design and optimize compute foundations for large-scale language model training. This role suits an engineer who thrives working close to the hardware and research boundary.

Role Overview

The Infrastructure Research Engineer position at Thinking Machines focuses on building, optimizing, and maintaining the compute foundations that power large-scale language model training. Professionals in this role develop high-performance machine learning kernels and improve the distributed compute stack. The position offers a chance to collaborate with researchers and systems architects to bridge algorithmic design with hardware efficiency, prototyping new kernel implementations and profiling performance across different hardware generations.

Key Responsibilities

Design and implement custom machine learning kernels including CUDA, CuTe, and Triton for core language model operations such as attention, matrix multiplication, gating, and normalization.
Engineer compute primitives aimed at reducing memory bandwidth bottlenecks and enhancing kernel compute efficiency.
Partner with research teams to ensure kernel-level optimizations align with overall model architecture and algorithmic goals.
Build and maintain a reusable library of kernels and performance benchmarks to support internal model training.
Support infrastructure stability, scalability, reproducibility, precision format consistency, and high utilization of compute resources.
Share technical insights through internal talks, technical papers, or open-source contributions.

Requirements

Bachelor’s degree or equivalent experience in computer science, electrical engineering, statistics, machine learning, physics, robotics, or a related field.
Strong software engineering abilities to write performant, maintainable code and debug complex codebases.
Solid understanding of deep learning frameworks like PyTorch and JAX alongside their underlying system architectures.
Ability to thrive in a collaborative environment with multiple cross-functional partners and subject matter experts.
A proactive bias for action to take initiative across different stacks and teams to ensure successful project delivery.
Proficiency in GPU programming frameworks such as CUDA, CuTe, or Triton.
Demonstrated capability in analyzing, profiling, and optimizing compute-intensive workloads.

Nice to Have

Experience training or supporting large-scale language models featuring tens of billions of parameters or more.
A track record of enhancing research productivity via infrastructure design or process improvements.
Experience developing or tuning kernels for deep learning frameworks or custom accelerators.
Familiarity with tensor parallelism, pipeline parallelism, or distributed data processing frameworks.
Experience implementing low-precision formats such as FP8, INT8, and block floating point, or contributing to related compiler stacks like XLA and TVM.
Previous contributions to open-source GPU, machine learning systems, or compiler optimization projects.
Prior research or engineering background in numerical optimization, communication-efficient training, or scalable artificial intelligence infrastructure.

Skills and Tools

CUDA
CuTe
Triton
PyTorch
JAX
Low-precision arithmetic (FP8, INT8, block floating point)
Distributed compute stacks and parallel strategies

Salary and Benefits

Expected annual salary range of $350,000 to $475,000 USD, determined by background, skills, and experience.
Comprehensive health, dental, and vision benefits.
Unlimited paid time off (PTO).
Paid parental leave.
Relocation support as needed.
Visa sponsorship is available for qualified candidates.

About the Company

Thinking Machines builds artificial intelligence designed to extend human will and judgment. The organization trains frontier models utilizing Inkling, creates Tinker to enable customized model creation, and crafts interfaces that expand human-AI communication.

How to Apply

Interested candidates should submit their application directly through the official Thinking Machines careers page.

Skills and tech stack

PyTorchLLM

Questions candidates ask

What is the salary range for the Research Engineer role?

The expected annual salary range is $350,000 to $475,000 USD, depending on background, skills, and experience.

Where is the position located?

The role is based in San Francisco, California, with remote options available.

Does Thinking Machines provide visa sponsorship?

Yes, Thinking Machines sponsors visas and works through the visa process together with the right candidate.

What programming frameworks and tools are required?

Candidates must be proficient in GPU programming frameworks like CUDA, CuTe, or Triton, and understand deep learning frameworks such as PyTorch or JAX.

What benefits are offered with this position?

Benefits include health, dental, and vision coverage, unlimited PTO, paid parental leave, and relocation support if needed.

Is this an evergreen job posting?

Yes, this is an evergreen role kept open on an ongoing basis to gather expressions of interest, and applications are continuously reviewed as new opportunities arise.

Ready to apply?

The application is on the official Thinking Machines careers page.

Apply on company site