Also, papers of the month will be taking a break. We will instead be focusing on a new blog series on a variety of topics, including technical deep dives, brief history of RL/quantisation/low bit formats, and maybe even some of our own exciting Ideas!
February Papers: Thinking Depth, Latent Actions, Quantization and Riemannian Flows. Last month saw another tsunami of great work, our team picked out 4 we'd like to share! graphcore-research.github.io/2026-03-03-p...
Check out the Graphcore Research team's efficiency-focused AI roundup: speeding up MoE models with SonicMoE, building fact-storing MLPs, and “byteifying” language models for better compute and character-level performance. graphcore-research.github.io/papers-of-th...
Austin, TX joins Graphcore (a SoftBank Group company) in building the future of AI compute. We've just shared the first engineering roles in the Lone Star State - with more to come. www.graphcore.ai/jobs
We’re marking #WorldQualityWeek 2025 alongside
The Chartered Quality Institute. If you're as passionate about the pursuit of excellence in industry as we are at Graphcore, come join us. We have vacancies across our Quality and Manufacturing Operations teams. job-boards.greenhouse.io/graphcore
It's World Quality Week 2025 - celebrating the pursuit of perfection and the processes that help businesses like Graphcore deliver the highest quality products and services. Our Head of Quality Kevin Newey reflects on this year's theme: Think Differently www.graphcore.ai/posts/thinki...
LLMs are eating the world.They've also been consuming the attention of Graphcore's research team as they pore over the past month's most significant papers.Check out their pick of LLM research from October, including 'Grouped Lattice Vector Quantisation'. graphcore-research.github.io/papers-of-th...
Graphcore is looking for world-class Characterisation and Bring-up Engineers to join our new Bengaluru AI Engineering Campus. Working with teams across the company, you’ll help bring to life the next generation of AI compute systems, as part of SoftBank Group job-boards.greenhouse.io/graphcore/jo...
Help us bring the world’s most advanced AI processors to life. Graphcore is hiring Post-Silicon Validation Engineers in Bengaluru. We have a range of roles at different levels. Find out more and apply: job-boards.greenhouse.io/graphcore/jo...
Join us at Graphcore’s new AI Engineering Campus in Bengaluru.
We’re hiring Silicon Verification Engineers to help us deliver an exciting new generation of AI compute hardware, as part of the SoftBank Group. Find out more: job-boards.greenhouse.io/graphcore/jo...
Graphcore is hiring Physical Design Engineers for our new AI Engineering Campus in Bengaluru. Graphcore is a SoftBank Group company so your work will be used by the world’s leading AI innovators. Find out more job-boards.greenhouse.io/graphcore/jo...
Graphcore has openings for Logical Design Engineers at our new AI Engineering Campus in Bengaluru. As part of the Silicon design team, you will be helping to
deliver leading-edge micro-architecture and RTL for a new generation of AI compute hardware. job-boards.greenhouse.io/graphcore/jo...
Graphcore's new AI Engineering Campus in Bengaluru is hiring. We're looking for semiconductor engineers to help build the future of AI compute as part of the SoftBank Group. Find out more, including how to apply: www.graphcore.ai/bengaluru
Namaskara Bengaluru! Graphcore – part of the SoftBank Group – is opening a new AI Engineering Campus. We're creating 100 silicon engineering jobs immedately, with hundreds more coming soon. Find out more: www.graphcore.ai/posts/graphc...
Large AI models and the compute systems they run on are inextricably linked. Graphcore is hiring for a Machine Learning Engineer (Large Systems) to help advance AI technology by developing and optimising AI models tailored to our specialised hardware. job-boards.greenhouse.io/graphcore/jo...
The Graphcore Research team's review of August's AI publications includes optimising multiple datasets with ADMIRE BayesOpt, fine-tuning 3D molecule generation diffusion models, and looks at Graph-R1’s approach to RAG. graphcore-research.github.io/papers-of-th...
How low can you go?
Graphcore's Alex Titterton explains how lower-precision number formats are enabling more efficient AI models, examines the quantization techniques that make low-precision AI possible, and asks whether this approach is suitable for more than just LLMs. youtu.be/Z9yHzALRLBk
Alex was speaking at the recent Bristol MLOps Meetup.
In his deep-dive talk - given at the recent MLOps Bristol Meetup - Graphcore's Alexandre Payot examines the techniques that made DeepSeek V3 possible, and how they can be applied to other Mixture of Experts models. youtu.be/iVUDqatG8lE?...
Thanks to everyone who came along to the MLOps Meetup at Graphcore HQ in Bristol. Great to discuss topics such as LLM efficiency, DeepSeek, using graphs in LLMs, and much more, with our rapidly growing community.
There's still time to sign up for this week's (March 6th) Bristol MLOps meetup. We'll be talking DeepSeek, LLM efficiency, and using graphs to produce naturalistic LLM chat systems. lu.ma/y3k6cr06
Silver Award mean is awarded to companies that ensur at least 95% of all invoices are paid within 60 days and at least 95% of small business invoices within 30 days.
Congratulations to Graphcore's Accounts Payable team on receiving a Silver Award in the Fair Payment Code Awards. Graphcore is also signed-up to the UK Small Business Commissioner's Fair Payment Code. www.smallbusinesscommissioner.gov.uk/new-fair-pay...
Join us for the next Bristol MLOps meetup on March 6th. We'll be talking DeepSeek, LLM efficiency, and using graphs to produce naturalistic LLM chat systems. Sign up via the link. lu.ma/y3k6cr06
DeepSeek is among the models getting the once-over by
@gcresearchteam.bsky.social in this month's Papers of the Month: www.graphcore.ai/posts/januar...
Join Graphcore's Director of AI Sylvain Viguier and other leading AI voices to discuss developments in open source artificial intelligence at OpenUK's State Of Open Con 25. London, 4th-5th Feb. stateofopencon.com
