<p>As a member of the Dojo Machine Learning software team, you will be responsible for developing and optimizing machine learning kernels to run on a massively parallel machine. The ideal candidate will have a strong background in kernel development and performance optimization for AI workloads, with a passion for delivering high-performance implementations, which are simple to use.</p> <ul> <li>Develop and validate datapath kernels on a massively parallel machine </li><li>Design and deliver implementations of known neural network algorithms on Dojo Hardware, for both training and inference workloads </li><li>Collaborate with architects and compiler engineers to transfer kernel designs to the Dojo compiler and production environment </li><li>Work closely with the team to drive improvements to the kernel development framework, supporting environments, and other components of the system </li><li>Work with hardware and simulations teams to ensure future hardware generations optimally satisfy algorithmic requirements </li><li>Participate in code reviews, testing, and debugging to ensure high-quality software </li><li>Stay up to date with the latest developments in AI workloads, domain-specific languages, computer architecture, and simulation techniques </li><li>Degree in Engineering, Computer Science, or equivalent in experience and evidence of exceptional ability </li><li>3+ years of experience in kernel development </li><li>In depth understanding of Large Language Models (LLMs), transformer architectures, their training, inference, and optimization </li><li>Strong programming skills in languages such as C/C++ and/or Python </li><li>Experience with deep learning frameworks such as PyTorch, JAX, Pallas etc. </li><li>Strong understanding of CPU and/or GPU microarchitecture, including pipelining, caching, and memory hierarchy </li><li>Excellent problem-solving skills, with the ability to analyze complex problems and develop creative solutions </li><li>Strong communication and collaboration skills, with the ability to work effectively with architects, engineers, and researchers </li><li>Able to work from Tesla's Palo Alto office </li></ul>