companyThinking Machines Lab logo

Infrastructure Research Engineer - Kernels at Thinking Machines | San Francisco

On-site Full-time $350K/yr - $475K/yr

Clicking Apply Now takes you to AutoApply where you can tailor your resume and apply.


Unlock Your Potential

Generate Job-Optimized Resume

One Click And Our AI Optimizes Your Resume to Match The Job Description.

Is Your Resume Optimized For This Role?

Find Out If You're Highlighting The Right Skills And Fix What's Missing

Experience Level

Experience

Qualifications

The ideal candidate will possess:A strong background in computer science, electrical engineering, or a related field. Proficiency in programming languages such as C++, Python, or similar. Experience with machine learning frameworks and tools. Familiarity with high-performance computing and GPU programming. A collaborative mindset and excellent communication skills.

About the job

At Thinking Machines Lab, our ambition is to enhance human potential by advancing collaborative general intelligence. We envision a future where individuals have the tools and knowledge to harness AI for their distinct requirements and aspirations.

Our team comprises dedicated scientists, engineers, and innovators who have contributed to some of the most renowned AI products, including ChatGPT and Character.ai, along with open-weight models like Mistral, and influential open-source projects such as PyTorch, OpenAI Gym, Fairseq, and Segment Anything.

About the Role

We are seeking an Infrastructure Research Engineer to architect, optimize, and sustain the computational frameworks that facilitate large-scale language model training. You will create high-performance machine learning kernels (e.g., CUDA, CuTe, Triton), enable effective low-precision arithmetic operations, and enhance the distributed computing infrastructure essential for training expansive models.

This position is ideal for an engineer who thrives in close collaboration with hardware and research disciplines. You will partner with researchers and systems architects to merge algorithmic design with hardware efficiency. Your responsibilities will include prototyping new kernel implementations, evaluating performance across various hardware generations, and helping to establish the numerical and parallelism strategies crucial for scaling next-generation AI systems.

Note: This is an evergreen role that remains open continuously for expressions of interest. We receive numerous applications, and there may not always be an immediate opportunity that aligns with your qualifications. However, we encourage you to apply, as we regularly assess applications and will reach out as new positions become available. You are also welcome to reapply after gaining additional experience, but please refrain from applying more than once every six months. Additionally, you may notice postings for specific roles catering to particular projects or team needs. In such cases, you are encouraged to apply directly alongside this evergreen listing.

What You’ll Do

  • Design and develop custom ML kernels (e.g., CUDA, CuTe, Triton) for key LLM operations such as attention, matrix multiplication, gating, and normalization, optimized for contemporary GPU and accelerator architectures.
  • Conceptualize compute primitives aimed at alleviating memory bandwidth bottlenecks and enhancing kernel compute efficiency.
  • Collaborate with research teams to synchronize kernel-level optimizations with model architecture and algorithmic objectives.
  • Create and maintain a library of reusable kernels and performance benchmarks that serve as the foundation for internal model training.
  • Contribute to the stability and scalability of our infrastructure, ensuring it meets the growing demands of AI development.

About Thinking Machines Lab

Thinking Machines Lab is on a mission to revolutionize human capability through the advancement of collaborative general intelligence. We strive to create accessible AI tools that empower individuals to realize their unique goals and aspirations.

Similar jobs

Tailoring 0 resumes

We'll move completed jobs to Ready to Apply automatically.