Search
Go
Research Scientist (Visual Generative AI & World Models)
GREATER LONDON · GB
—
Full-time
SENIOR
About the Role
Graphcore is seeking a Research Scientist to advance AI research in visual generative modeling, multimodal learning, world models, and hardware-aware machine learning. The role involves exploring new architectures and training methods for applications in embodied AI and robotics, and collaborating with silicon teams to influence future AI compute systems.
Responsibilities
- Develop and evaluate new ideas in visual generative AI, multimodal modeling, and world models from hypothesis through publication.
- Prepare, submit, and present work at AI conferences.
- Collaborate with researchers and engineers to understand how emerging AI workloads shape Graphcore hardware and software.
Requirements
- Master's, PhD, or equivalent in a technical discipline.
- Experience in visual generative AI, visual understanding, or world models.
- Strong Python skills with PyTorch or JAX.
- Familiarity with deep learning fundamentals and experiment design.
- Evidence of research ability via publications or projects.
Nice to Have
- Experience with multimodal reasoning, embodied AI, or robotics.
- Lower-level programming (C++/CUDA/Triton).
- Knowledge of hardware considerations for deep learning (parallelism, memory hierarchy, etc.).
Graphcore
Bristol · GB · 552+ employees
Graphcore is a semiconductor company that develops accelerators for AI and machine learning, specifically its proprietary Intelligence Processing Unit (IPU) designed for massively parallel processing of machine learning models. The company, now a subsidiary of SoftBank Group, focuses on enabling breakthroughs in machine intelligence.
