Adam Jelley

Research / Co-Founder

General Intuition

Hello!

I’m a Member of Technical Staff / Co-Founder at General Intuition. We work on foundation models that require deep spatial and temporal reasoning. Our models are built on the billions of hours of gameplay recorded with Medal, with applications to the video game industry and beyond.

Before starting General Intuition, I was a PhD student working on efficient reinforcement learning approaches and world models in the Bayesian and Neural Systems group at the University of Edinburgh, lucky to be supervised by Professor Amos Storkey in the School of Informatics as well as Sam Devlin from the Game Intelligence Team at Microsoft Research Cambridge. I was supported by a Microsoft Research Scholarship.

Interests

Reinforcement Learning
Imitation Learning
World Models
RL from Human Feedback
Self-Supervised Learning
AI for Science

Education

PhD in Generating Environments and Pre-Training Agents for Efficient Reinforcement Learning, 2025
University of Edinburgh
MSci in Theoretical Physics, 2016
University of Cambridge
MA in Natural Sciences - Physics, 2015
University of Cambridge

Selected Publications

Diffusion for World Modeling: Visual Details Matter in Atari

We introduce DIAMOND, an reinforcement learning agent trained in a diffusion world model. Presented at NeurIPS 2024 (Spotlight).

Eloi Alonso, Adam Jelley, Vincent Micheli, Anssi Kanervisto, Amos Storkey, Tim Pearce, François Fleuret

Diffusion for World Modeling: Visual Details Matter in Atari

Aligning Agents like Large Language Models

An investigation into training agents like Large Language Models (LLMs) by unsupervised pre-training, supervised fine-tuning, and finally reinforcement learning from human feedback (RLHF). Presented at RLBRew Workshop at RLC 2024.

Adam Jelley, Yuhan Cao, Dave Bignell, Sam Devlin, Tabish Rashid

Aligning Agents like Large Language Models

Efficient Offline Reinforcement Learning: The Critic is Critical

An approach for efficient offline reinforcement learning by first learning the behaviour policy and values with supervised learning, before improving on this policy with reinforcement learning. Presented at ARLET Workshop at ICML 2024.

Adam Jelley, Trevor McInroe, Sam Devlin, Amos Storkey

Efficient Offline Reinforcement Learning: The Critic is Critical

Contrastive Meta-Learning for Partially Observable Few-Shot Learning

An approach for meta-learning contrastive representations under partial observability. We demonstrate this approach can be utilised by reinforcement learning agents to learn a representation of their environment. Presented at ICLR 2023.

Adam Jelley, Amos Storkey, Antreas Antoniou, Sam Devlin

Contrastive Meta-Learning for Partially Observable Few-Shot Learning

Deep Ocean Learning of Small Scale Turbulence

We show machine learning can be successfully employed to infer turbulent mixing from quantities measured routinely by global observational programs. Published in GRL 2022.

Ali Mashayek, Nick Reynard, Fangming Zhai, Kaushik Srinivasan, Adam Jelley, Alberto Naveira Garabato, Colm-cille P. Caulfield

Deep Ocean Learning of Small Scale Turbulence

Experience

Research / Co-Founder

General Intuition

Apr 2025 – Present

Research into foundation models that require deep spatial and temporal reasoning.

PhD Candidate

University of Edinburgh

May 2021 – Jul 2025 Edinburgh

PhD Candidate in Efficient Reinforcement Learning and World Models.

Thesis: Generating Environments and Pre-Training Agents for Efficient Reinforcement Learning

Research Scientist Intern

Microsoft Research Cambridge

Jun 2023 – Sep 2023 Cambridge

Developed a pipeline for aligning agents with preferences on the Xbox game Bleeding Edge, for research into capabilities and limitations of Reinforcement Learning from Human Feedback (RLHF) in this domain.