Teams at Graphcore

Develop and optimise state-of-the-art machine learning models for specialised hardware and systems containing thousands of accelerators. Responsibilities include model training, benchmarking, profiling, kernel optimisation, experiment design, software testing, code reviews, and collaboration with research, software, and product teams to advance AI hardware and software. The role also involves distributed training and inference, performance analysis, and engagement with the broader AI community.

Develop, train, benchmark, and optimize state-of-the-art machine learning models across systems with thousands of accelerators. Identify performance bottlenecks, optimize kernels and software, conduct AI experiments, evaluate internal releases, and collaborate with research, software, product, and hardware teams. The role requires expertise in deep learning, hardware acceleration, performance optimization, and Python or C++ development, with desirable experience in distributed training, MLOps, LLMs, and HPC technologies.

Develop and optimize state-of-the-art machine learning models for specialized AI hardware and systems containing thousands of accelerators. Responsibilities include model training, benchmarking, profiling, kernel optimization, experimentation, software testing, code reviews, and collaboration with research, software, and product teams. The role focuses on performance, scalability, distributed training and inference, and advancing next-generation AI hardware and software.