Research Engineer, World Models, DeepMind
London, UK
Full Time
Role Summary
This role involves implementing core infrastructure and conducting research to build generative models of the physical world. The individual will solve problems to train world simulators, develop metrics for physical intelligence, curate and annotate training data, enable real‑time interactive generation, and develop methods for systems and infrastructure.
Role Highlights
* Python
* C++
* AI
* NLP
* Data Mining
* Training Data
* Research
* PhD
* Embedded
* ProtoTyping
* Compilers
* Search
* Infrastructure
* Inference
* pre‑training
Tools, Libraries and Frameworks
* JAX
* GPUs
* PyTorch
* NumPy
* Matplotlib
Responsibilities
* Implement core infrastructure for training world simulators.
* Conduct research on generative models of the physical world.
* Develop metrics for physical intelligence.
* Curate and annotate training data.
* Enable real‑time interactive generation.
* Develop systems and infrastructure methods.
Required Qualifications and Skills
* Bachelor's degree or equivalent practical experience and at least two years of software development experience using Python or C++.
* Experience with deep learning frameworks such as JAX or PyTorch.
* Proficiency in scientific computing libraries like NumPy, Pandas, and Matplotlib.
* PhD with a research focus on Machine Learning, Reinforcement Learning, or Multi‑Agent Systems preferred.
* Experience training models on accelerators in a distributed environment.
* Expertise in building and optimizing complex systems in JAX.
Disclaimer: Job and company description information and some of the data fields may have been generated via GPT‑4 summarisation and could contain inaccuracies. The full external job listing link should always be relied on for authoritative information.
#J-18808-Ljbffr