Roussel.

PhD student @ University of Bristol

“Give me a lever long enough and a fulcrum on which to place it, and I shall move the world.” ― Archimedes

ddrous@web:~

Selected Stories*

*Stay up to date by signing up for my newsletter.

Linear Attention in JAX Exploring efficient attention mechanisms through JAX implementation Coming soon

A Parallel Theory of the Causal Mind Investigating how our minds process causality through parallel mechanisms

Welcome to My Brand-New Website! An introduction to this space where I share research insights, technical explorations, and thoughts on AI

Mesh-Free Simulation in the Age of Big Data How modern mesh-free methods are revolutionizing computational physics and engineering Coming soon

Selected Publications*

*Please check my Google Scholar for more.

AAAI 2026 Student Abstract

Language Models Do Not Embed Numbers Continuously

A Davies, RD Nzoyem, N Ajmeri, T Silva Filho

ArXiv 2025

Weight-Space Linear Recurrent Neural Networks

RD Nzoyem, N Keshtmand, E Crespo Fernandez, I Tsayem, R Santos-Rodriguez, DAW Barton, T Deakin

SCOPE@ICLR 2025

Towards Foundational Models for Dynamical System Reconstruction: Hierarchical Meta-Learning via Mixture of Experts

RD Nzoyem, G Stevens, A Sahota, DAW Barton, T Deakin

CoLLAs 2025

Reevaluating Meta-Learning Optimization Algorithms Through Contextual Self-Modulation

RD Nzoyem, DAW Barton, T Deakin

ICLR 2025

Neural Context Flows for Meta-Learning of Dynamical Systems

RD Nzoyem, DAW Barton, T Deakin

SC’W 2023

A Comparison of Mesh-Free Differentiable Programming and Data-Driven Strategies for Optimal Control

RD Nzoyem, DAW Barton, T Deakin

Recent Services

Reviewer:

Euro-PAR’24 ICLR’25 ICLR’26 ICML’25 (Top Reviewer) NeurIPS’25 SCOPE@ICLR 2025 TMLR JAIR

Conferences:

The Interactive AI Spring Research Conference Co-organizer | March 2024, Bristol, UK

Recent Talks

Weight-Space Linear Recurrent Neural Networks Workshop on Scientific Machine Learning, University of Bristol, UK | June 2025

Engineering Mathematics at the University of Bristol STEM Up Virtual Programme, Outreach & Widening Participation Team | May 2025

My Background, Research, Outreach & Gender Equality in STEM Engineering Design + Black Engineers Talk Series, University of Bristol, UK | Feb 2025

Neural Context Flows for Meta-Learning of Dynamical Systems EPS Seminar Series, University of Bristol, UK | Feb 2025

Quarto: The Tech and its Uses New Ways of Publishing Workshop, University of Bristol, UK | Jun 2024

Differentiable Programming for Mesh-Free Fluid Control Physics Enhancing ML in Applied Mechanics, Institute of Physics, London | Nov 2023

User-centered PDE solvers for optimal control and reconstruction Pecha Kucha Student Talk, Spring Research Showcase, UK | Mar 2024

Scaling Differential Equations via Transferable Dynamics Flash Presentation, Barton Research Group Meeting, UK | Dec 2023

Emerging Techniques and Applications of Graph Neural Networks Graduate Student Seminar, CMU Africa, Kigali, Rwanda | Mar 2023

Latest News*

24 Jun. 2025: I am co-organizing a two-days conference focused on embedding domain-specific knowledge in ML models, physical and mathematical methods for AL/ML development, and applications. ☞ Programme

10 Jun. 2025: I’m giving a flash talk the UoB’s SciML in the Faculty of Engineering Workshop on a new adaptation rule I am extremely excited about: Weight-Space Linear Recurrent Neural Networks. ☞ Slides

19 May. 2025: I was in Montreal for the CRM summer school on the Mathematical Foundations of Data Science. I also attended the workshop on Optimization and learning: theory and applications !

22 Apr. 2025: I’m attending ICLR’25 in Singapore and presenting lots of stuff. This time again, I’m looking to connect and have as much fun as possible. ☞ LinkedIn Post

06 Mar. 2025: Our latest paper MixER: Better Mixture of Experts Routing for Hierarchical Meta-Learning was unanimously accepted into the SCOPE workshop @ ICLR 2025. Looking forward to seeing everyone in Singapore !

13 Feb. 2025: I gave a talk at the University of Bristol’s EPS Seminar Series and the Engineering Design Society. It’s about Neural Context Flows and follow-up work on foundational models for science. ☞ Slides

22 Jan. 2025: I’m delighted to announce that our paper Neural Context Flows has been accepted to ICLR 2025. More info to come, including a blogpost with code and weights !

07 May. 2024: I’m attending ICLR’24 in Vienna and presenting a poster at the workshop on AI4DifferentialEquations in Science.

17 Apr. 2024: My profile now appears on the Engineering Includes Me wall in the University of Bristol’s iconic Queen’s Building. ☞ Blog Post

13 Mar. 2024: I’m going to the Turing Institute to attend the launch of their Probabilistic Programming theme.

20 Nov. 2023 I’m giving a talk at the 2nd workshop on Physics Enhancing Machine Learning in Applied Mechanics, held at the Institute of Physics in London, UK. ☞ Slides.

13 Nov. 2023: Our paper on mesh-free differential programming for optimal control is out during SuperComputing’23. ☞ Read here.

18 Sep. 2023: I’m attending the 175th European Study Group with Industry (ESGI) in Berlin, Germany.

01 Aug. 2023: I’ve pre-released version 0.1.4 of Updes: a mesh-free differentiable software for quick experimentation with a variety of PDEs. ☞ Repo.

10 Mar. 2023 I gave a talk at CMU Africa on emerging techniques and applications of graph neural networks. ☞ Recording.

01 Dec. 2022: I began my research on AI, HPC, and Scientific Machine Learning for mesh-free simulations.

10 Oct. 2021: I moved to Bristol for a PhD in the Interactive AI CDT.

Reach out