Senior Machine Learning Engineer (Large Systems)

About Graphcore

Graphcore is at the forefront of AI computation, assembling a team of semiconductor, software, and AI specialists with extensive expertise in delivering complete AI computing solutions, from silicon design to software and large-scale data center infrastructure. Supported by substantial investment from the SoftBank Group and integrated within the SoftBank AI ecosystem, Graphcore is expanding its global teams to address the transformative potential of AI technology.

Job Summary

We seek a Senior Machine Learning Engineer for our Applied AI team to drive advancements in AI by developing and optimizing AI models specifically for our unique hardware. This role focuses on large-scale systems where optimizing performance is crucial. The successful candidate will collaborate closely with software development and research teams to innovate and enhance Graphcore's AI technology offerings.

Responsibilities

• Develop and optimize state-of-the-art machine learning models for high accuracy and performance across thousands of accelerators.
• Evaluate new internal software releases, providing constructive feedback, implementing fixes, and performing code reviews.
• Benchmark machine learning models to identify and address performance limitations, enhancing efficiency.
• Design, conduct, and analyze experiments on novel AI techniques, integrating findings into implementations.
• Work collaboratively with Research, Software, and Product teams on next-generation AI hardware developments.
• Maintain engagement with the AI community to stay informed about the latest trends and advancements.

Candidate Profile

Essential Qualifications and Skills:

• Academic background at Bachelor, Master’s, or PhD level in Machine Learning, Computer Science, Mathematics, Data Science, or related discipline, or equivalent experience.
• Proficiency with deep learning frameworks such as PyTorch and JAX.
• Strong programming experience in Python or C++.
• Expertise covering the full cycle of deep learning including training, optimization, and evaluation.
• Ability to design, execute, and report on machine learning experiments effectively.
• In-depth understanding of performance bottlenecks and skill in addressing them.
• Adaptability and the ability to work efficiently in a fast-paced, dynamic environment.
• Collaborative mindset comfortable with cross-team cooperation.
• Excellent communication skills for explaining complex technical concepts to diverse audiences.

Desirable Skills and Experience:

• Experience with MLOps frameworks on Kubernetes clusters.
• Building scalable production systems utilizing large language models.
• Expertise in efficient computing through low-precision arithmetic techniques.
• Experience developing performance-optimized kernels with C++, Triton, or CUDA.
• Experience in distributed training or inference over clusters of 64 or more accelerators.
• Knowledge of high-performance computing systems and networking technologies like Infiniband, NVLink, and RoCE.
• Contributions to open source projects or relevant published research.
• Familiarity with cloud computing platforms.
• Enthusiasm for sharing knowledge through presentations, publications, and talks within the AI community.

Benefits and Additional Information

Graphcore provides a competitive salary, flexible working options, generous annual leave, comprehensive private medical and dental insurance, pension matching up to 5%, life assurance, and income protection. The company supports parental leave, employee assistance programs including mental health and bereavement support, healthy food provision, and an inclusive and diverse work culture that respects all differences. Reasonable adjustments are available throughout the recruitment process.

Applicants must have the legal right to work in the UK; currently, Graphcore does not offer visa sponsorship or support with visa applications.