Rosie Wood

Research Software Engineer · Alan Turing Institute

I am a Research Software Engineer working in the Research Engineering Group and Research Computing team at the Alan Turing Institute. From 2023 to 2025 I worked on MapReader — a computer vision pipeline for analysing large collections of historical maps — as part of the Living with Machines and Data/Culture projects, and remain a core maintainer of the software. Since then, I’ve been part of the t0 team, where I work on applications of local LLMs, and the Research Computing team, where I help manage Turing’s HPC allocations, support researchers in running workloads on GPU and organise/run HPC-related trainings. I also co-lead the REG Open Source Service Area and sit on the AIRR User Advisory Group.

I have a background in chemistry (MSci, University of Bristol) and computational chemistry (MRes, University of Edinburgh).

GitHub LinkedIn ORCID

Projects

Default Most Recent

MapReader

2023–present

Award-winning computer vision pipeline for large-scale analysis of historical maps.

t0-1

2025–2026

A RAG system which uses small language models fine-tuned for reasoning to answer domain specific queries (focused on NHS A-Z conditions).

Eden

2026–present

A RAG system which uses small reasoning models to answer domain-specific queries about gardening (focused on RHS data).

multi_node_inference

2025

A series of scripts for running multi-node inference with LLMs across multiple nodes on HPC using vLLM (including containerised workflows).

reginald

2024

A slack bot/chat interface which uses open source LLMs and RAG to answer queries about Turing internal data.

Talks & Presentations

Recognition

  • Featured in SC25 HPC Ignites ‘Women in HPC’ series. 03/2025
  • MapReader awarded the 2023 Roy Rosenzweig Prize for Creativity in Digital History. 12/2023