Research Scientist, Graphcore
Visual Generative AI & World Models
2+ days a week in office
Advanced AI systems and services
Be an early applicant
Advanced AI systems and services
501-1000 employees
Be an early applicant
2+ days a week in office
501-1000 employees
Company mission
To make machine learning faster, easier and more intelligent.
Role
Who you are
- Master’s, PhD or equivalent experience in a technical discipline (e.g., Mathematics, Statistics, Computer Science, Physics, Chemistry, Biomedical Engineering)
- Experience in visual generative AI, visual understanding or world models
- Strong Python programming skills using a modern deep learning framework, e.g. PyTorch or JAX
- Familiarity with deep learning fundamentals, including model architectures, optimisation, evaluation and scaling
- Ability to design, execute, analyse and clearly communicate ML experiments
- Mathematical foundations to support the above, including calculus, probability theory and linear algebra
- Evidence of research ability, such as conference or workshop submissions, publications, technical reports, open-source projects or impactful industrial research
Desirable
- Experience with multimodal reasoning or generation, action-conditioned models, embodied AI, robotics or autonomous systems
- Lower-level programming for hardware efficiency, e.g. C++/CUDA/Triton
- Practical familiarity with hardware considerations for deep learning, such as parallelism, memory hierarchy, vector and matrix engines, data movement, bandwidth limits and performance bottlenecks
- Practical familiarity with deep learning software stacks, such as graph compilation, kernel fusion, XLA/ATen operations, streams and asynchronous execution
What the job involves
- As a Research Scientist at Graphcore, you will advance AI research at the intersection of visual generative modelling, multimodal learning, world models and hardware-aware machine learning.
- You will explore new model architectures, training methods and deployment strategies with applications in embodied AI, robotics and autonomous systems. Example research directions could include efficient video generation, diffusion and flow-based models, multimodal representation learning, world models for agents, or analysis of how emerging generative AI workloads influence future AI accelerators
- This role sits at the interface between frontier model research and AI hardware. Specialised hardware has been a key driver of AI progress over the last decade, and we believe that hardware-aware AI algorithms and AI-aware hardware developments will continue to be critical to advancing this field.
- We are looking for researchers and engineers with the theoretical depth, practical judgement and implementation skills to turn ambitious ideas into rigorous experiments, publications and technical insights that influence the future of AI compute
- Develop and evaluate new ideas in visual generative AI, multimodal modelling and world models, from initial hypothesis through experiment design, implementation, analysis and publication
- Prepare, submit and present your work to AI conferences and workshops
- Work with researchers, software engineers and silicon teams to understand how emerging AI workloads can shape, and be shaped by, future Graphcore hardware and software systems
Salary benchmarks
Our take
The hardware required to train AI models comes at a high cost, both financially and environmentally, producing millions of dollars in expenses and hundreds of tons of greenhouse gases. In response, Graphcore develops accelerator chips designed to expedite AI system training while significantly reducing power consumption.
AI accelerators like those crafted by Graphcore specialise in enhancing specific AI applications, such as artificial neural networks, deep learning, and machine learning. Utilising multicore designs and prioritising low-precision arithmetic or in-memory computing, these accelerators elevate the performance of large AI algorithms, achieving breakthrough results in natural language processing and computer vision.
Graphcore has ambitious plans to create an AI supercomputer, aiming to be one of the most powerful in its class. However, progress on this has extended beyond its planned 2024 debut, and the company was purchased by Softbank later that year. Though the acquisition by the Japanese multinational lowered Graphcore's valuation and highlighted structural problems in the local financing of UK tech pioneers, the buyout will undoubtedly aid the company in its ambitions to shape the future of AI computing.
Freddie
Company Specialist at Welcome to the Jungle
Insights
Company
Funding (last 2 of 6 rounds)
Dec 2020
$222m
SERIES E
Feb 2020
$150m
SERIES D
Total funding: $682m
Company benefits
- Work-Life Balance: We believe you should have a good work-life balance. Everyone's situation is different, so we offer flexible working arrangements and start and finish times, allowing you to choose when to work from home or come into the office
- Wellbeing: We want to ensure that our employees have the resources and support they need to prioritise their health and wellbeing
- Healthcare: We offer Private Medical Insurance (PMI) to our employees to help them take care of their health and that of their loved ones
- Progression: We prioritise learning and development as an integral part of our culture at Graphcore. Our Talent Development team is dedicated to equipping all our employees with the necessary tools, resources, and support to enhance their skills and reach their full potential in their careers
- Pension: We believe that our employees deserve the best when it comes to planning for their future
- Income Protection: Graphcore understands that unforeseen circumstances can happen, and we want to ensure our employees have the necessary financial security in place
Company HQ
Old City, Bristol, UK
Leadership
Nigel Toon
(Co-Founder & Exec Chair)Previously the President and CEO at picoChip Designs (processors for telecoms). As well as the VP and GM of Europe at Altera (semiconductors, merged with Intel).
Simon Knowles
(Co-Founder & CTO)Founded two successful semiconductor companies, Element14 and Icera, for a combined value of over $1 billion.
Share this job
View 79 more jobs at Graphcore