Research Scientist, Graphcore
2+ days a week in office
Advanced AI systems and services
Open for applications
Advanced AI systems and services
201-500 employees
Open for applications
2+ days a week in office
201-500 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 field (e.g., Maths, Statistics, Computer Science, Physics, Chemistry, Biomedical Engineering)
- Python programming in a modern deep learning framework, e.g. PyTorch or JAX
- Familiar with deep learning fundamentals: models, optimisation, evaluation and scaling
- Capable of designing, executing and reporting from ML experiments
- Mathematics skills to support the above: calculus, probability theory and linear algebra
Desirable
- Experience in one or more of: {distributed computing, efficient computing based on low-precision arithmetic, deep learning models including large generative models for language, vision and other modalities, machine learning for molecules and proteins (ideally with some background in chemistry and biological sciences)}
- Lower-level programming for hardware efficiency, e.g. C++/CUDA/Triton
- Practical familiarity with hardware capabilities for deep learning – threads, caches, vector & matrix engines, data dependencies, bus widths and throttling
- Practical familiarity with software stacks for deep learning – compilation, kernel fusion, XLA/ATen ops, streams, and asynchronous execution
- Experience submitting papers to international scientific conferences or workshops
What the job involves
- As a researcher at Graphcore, you will contribute to the advancement of AI research, investigating new ideas that push the limits on important AI/ML problems
- Specialised hardware has been the key driver of the progress of AI over the last decade, and we believe that hardware-aware AI algorithms and AI-aware hardware developments will continue to be critical to the advancement of this exciting field
- Generate AI/ML ideas, design experiments, implement them & evaluate results
- Prepare, submit & present your work to AI conferences and workshops
- Collaborate with researchers, silicon and software engineers at Graphcore to help define, build and test Graphcore’s next generation of AI hardware
- Graphcore Research participates in both fundamental and applied research, to characterise the computational requirements of machine intelligence and to demonstrate how hardware can drive the next generation of innovative AI models
- We publish at leading AI/ML conferences (NeurIPS, ICML, ICLR) as well as specialist workshops, and collaborate with other research teams and organisations across the world
Salary benchmarks
Our take
Traditional hardware, mainly GPUs, used for training these 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 is developing 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 creat an AI supercomputer set to debut in 2024, aiming to be one of the most powerful in its class. The company is actively seeking strategic partnerships to bring this vision to fruition, signaling its commitment to shaping 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
Articles
Leadership
Nigel Toon
(CEO)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
(CTO)Founded two successful semiconductor companies, Element14 and Icera, for a combined value of over $1 billion.
People progressing
Joined Graphcore as a Test Technician in 2021 and has since been promoted to Semiconductor Test Engineer in 2019 and Staff SIlicon Test Engineer in 2023.
Share this job
View 1 more job at Graphcore