AI Performance Engineer
Graphcore is seeking an AI Performance Engineer to join our System Engineering Performance team in Milpitas, California. This team is dedicated to architecting and optimizing high-performance infrastructure for large-scale datacenter deployments, ensuring efficiency, scalability, and reliability across the full hardware-software stack. As part of Graphcore, a leading innovator in artificial intelligence compute, you'll contribute to developing hardware, software, and systems infrastructure that unlocks the next generation of AI breakthroughs.
In this role, you will analyze machine learning models' compute and memory requirements using roofline analysis and simulations. Collaborating across hardware and software teams, you'll optimize large-scale AI workloads, benchmark and monitor system performance across distributed systems, and troubleshoot any issues that arise. Additionally, you'll optimize communication stacks, including MPI, NCCL, UCX, RDMA, and networking fabrics, and profile AI workloads to identify and address performance bottlenecks. Developing high-quality, ARM-compatible code and documentation will also be a key part of your responsibilities.
The ideal candidate will possess a BS or MS in Computer Science, Electrical Engineering, or a related field. Essential qualifications include experience with distributed systems and communication libraries such as MPI, NCCL, UCX, or libfabric, strong programming skills in C++ and Python, and experience profiling and optimizing HPC or AI/ML workloads. Familiarity with ML benchmarks like MLPerf is also required. Desirable qualifications include experience with GPUs or accelerated computing architectures, knowledge of HPC networking and interconnect technologies such as InfiniBand or RoCE, familiarity with ML frameworks like PyTorch or TensorFlow, understanding of ARM architectures and toolchains, and strong debugging, profiling, and performance optimization skills.
Joining Graphcore offers the opportunity to work in a culture of continuous learning and constant innovation, alongside a diverse team of AI research specialists, silicon designers, software engineers, and systems architects. As part of the SoftBank Group, Graphcore is committed to enabling Artificial Super Intelligence and ensuring its benefits are accessible to everyone. This role provides a chance to contribute to cutting-edge AI solutions and advance your career in a dynamic and forward-thinking environment.