About

I'm a Senior Research Associate in Machine Learning in Oxford's Department of Statistics, funded by the Leverhulme Trust and supervised by Prof Yee Whye Teh, Prof Nathalie Seddon and Dr Steven Reece. I use large language models, NLP and information retrieval to mine the evidence base for nature-based solutions to climate change, with the Leverhulme Centre for Nature Recovery and the Nature-based Solutions Initiative. It's also a vantage point onto other projects — mainly LLM robustness and its real-world applications.

Before Oxford, I worked across computational sciences: comparative genomics in Switzerland on the OMA project; a PhD in computational neuroscience at Imperial — during which I also worked with BarefootLaw on Winnie, an NLP tool widening access to justice across East Africa; and a research residency at X, Alphabet's moonshot factory, probing frontier LLMs and helping start its moonshot for professional intelligence.

If there's a through-line, it's the kind of problem I'm drawn to: wicked ones — messy, cross-disciplinary, resistant to any single method. I try to fall in love with the problem, not the solution — and follow it wherever it leads.

Increasingly, that pulls me toward mission-driven applied ML — agents for science, and AI for climate and the environment — and toward teams building toward those goals at scale.

Research interests

Large language models and NLP — their evaluation, robustness and real-world applications; AI agents and evidence synthesis for the scientific process, especially nature-based solutions and climate; and neuro-AI, taking learning inspiration from the brain.

Projects

2026
Co-founded and organised a 22-person hackathon across Oxford / NUS / NTU. Shipped five open-source agents for academic workflows. Co-architected Prior: an auditable, agentic literature-contribution graph over research fields — provenance, cross-paper typed relations, and surfaced contradictions — with an IPCC-inspired framework for evidence calibration in development. Sponsored by Anthropic.
2026
Working-group member of an ESIIL-led initiative developing best practices and tooling for AI-assisted text analysis in environmental research and policy.

Education

2019–2024
PhD, Computational Neuroscience · Imperial College London
Supervised by Prof Claudia Clopath · funded by the Wellcome Trust. Thesis on biologically-plausible learning: network homeostasis, robust motor control, and credit assignment through space and time.
2018–2019
MSc, Theoretical Systems Biology & Bioinformatics · Imperial College London
Funded by the Wellcome Trust. Three research projects: self-organising systems with Prof Robert Endres; multi-scale clustering of single-cell RNA-seq data via Markov Stability with Prof Mauricio Barahona; and an inhibitory learning rule for network homeostasis with Prof Claudia Clopath, which became Kaleb et al. 2021.
2015–2018
BSc, Biomedical Sciences · UCL
Final-year research at the Francis Crick Institute with Sir Peter Ratcliffe and Prof Chris Barnes, on kidney-cancer transcriptomics (bulk RNA-seq). Earlier, an internship in comparative genomics in Switzerland with Prof Christophe Dessimoz and Dr Adrian Altenhoff on the OMA project.

Patents

US Patent Application × 3 (co-inventor, Google X).

2026
US Patent Application 18/785,645 · co-inventor.
2024
US Patent Application 18/674,444 · co-inventor.
2024
US Patent Application 18/589,228 · co-inventor.

Up-to-date list on Google Scholar.

Publications & writing

2025
Workshop on Reasoning & Planning for LLMs, 2025.
2024
NeurIPS (Advances in Neural Information Processing Systems 37), 2024.

Full list: Google Scholar · ORCID · ResearchGate.

Talks & service

Presented posters at COSYNE (2020, 2022) and NeurIPS (2024); reviewer for ICLR workshops (2023) and NeurIPS (2026).

Teaching

2025, 2026
Guest lecturer — Large Language Models
Lectures on LLMs for Oxford's Intelligent Earth and Big Data Institute (“AI for Healthcare”) Centres for Doctoral Training.
2020–2022
Graduate Teaching Assistant · Imperial College London
Teaching assistant for the Computational Neuroscience course, Department of Bioengineering.

Awards

2024
US visa, in the field of computational sciences.
2018
Wellcome Trust PhD studentship
4-year programme in Theoretical Systems Biology & Bioinformatics, Imperial College London.
2014
Highest mark in Kenya, Cambridge International AS-Level Physics.

Off the clock

I grew up in East Africa with front-row access to some of the world's most incredible mountains — I've climbed Mount Stanley, Mount Kenya, Mount Napak and Mount Sabyinyo. That upbringing left me with a global conscience — and is a big part of why I work on the environment today.

Sunrise at the peak of Mount Napak, Uganda, 2022
Sunrise at the peak of Mt Napak in 2022.

Contact

Best reached via my Oxford page, LinkedIn, Bluesky or X — or email me.