1 month ago
Bristol, United KingdomSenior
Responsibilities
- Implement and train state-of-the-art machine learning models and optimize them for performance, accuracy, and scalability across systems with thousands of accelerators.
- Benchmark and profile machine learning models to identify performance bottlenecks.
- Develop understanding across the software stack and optimize kernel implementations.
- Test and evaluate internal software releases, provide feedback, make code fixes, and conduct code reviews.
- Design, execute, and evaluate experiments involving novel AI methods.
- Collaborate with Research, Software, and Product teams to define, build, and test next-generation AI hardware.
- Engage with the AI community and stay current with developments in AI.
Requirements
- Bachelor’s, master’s, PhD, or equivalent experience in machine learning, computer science, mathematics, data science, or a related field.
- Proficiency with deep learning frameworks such as PyTorch or JAX.
- Strong software development skills in Python or C++.
- Expertise in hardware-accelerated deep learning from model training through optimization and evaluation.
- Ability to design, execute, and report on machine learning experiments.
- Understanding of performance bottlenecks and how to address them.
- Desirable experience with Kubernetes-based MLOps, production large language model systems, low-precision arithmetic, C++/Triton/CUDA kernels, distributed training or inference across 64 or more accelerators, HPC systems and networking, cloud computing platforms, open-source projects, or relevant research publications.
Benefits
- Hybrid working model from the Bristol, London, or Cambridge offices.
- Flexible working arrangements and generous leave.
- Up to 5% matched pension contribution, phantom equity, income protection, and life assurance.
- Designed office spaces with free food and drinks.
- Electric vehicle scheme through salary sacrifice.
- Flexible benefits including private medical insurance, dental cover, health cash plan, cycle-to-work scheme, and give-as-you-earn.
- Interview adjustments are available for candidates who need them.
Tech Stack
Categories
About Graphcore
Graphcore designs and sells Intelligence Processing Units (IPUs), systems, and a full software stack for training and inference of AI models in data centers and research labs. Revenue comes from hardware systems and associated software tools and services, used for workloads across NLP, computer vision, and graph neural networks. Founded in 2016 and headquartered in Bristol, the company is part of SoftBank Group and focuses on on‑prem and cloud deployments.