Evaluating LLMs for Responsible Scientific Data Rescue — Shona Ferguson | AI in Production 2026
Can LLMs accelerate scientific data rescue without losing accuracy or provenance? A head-to-head evaluation on historical cloud and rain chemistry data. Presented by Shona Ferguson (UK Centre for Ecology and Hydrology) at AI in Production 2026 — The Catalyst, Newcastle, 5 June 2026. ABOUT THE TALK Data are vitally important outputs from research, and, alongside answering research questions, they can provide opportunities for new research topics and hypotheses. There is a wealth of inaccessible historical data resources held by research institutions in proprietary and poorly documented formats. Preserving and making these datasets accessible through data rescue is crucial to prevent loss of the data and to support future re-use. However, manual data rescue can be a laborious and time-consuming task. Recent advancements in large language models (LLMs) present an opportunity to increase efficiency in data rescue workflows, yet their suitability for handling scientific tabular data remains under-explored. This work compares the performance of three commonly used LLMs and evaluates their performance in accelerating a digital data rescue methodology applied to historical cloud and rain chemistry data. The LLMs are compared across a series of structured tasks including prioritisation of data rescue, variable and unit identification, and metadata extraction. Their performance is assessed against a fully manual approach to identify the most significant efficiency gains and to examine risks related to hallucinations, inconsistencies, and loss of scientific context. Although LLMs can substantially reduce manual effort in data rescue tasks, it is vital to maintain a level of human quality control to ensure accuracy, provenance and reproducibility of important scientific data. This talk provides practical lessons for applying LLMs to real-world scientific data processing, contributing to broader discussions on evaluation, trust, and reliability of foundation models beyond natural language tasks. SPEAKER Shona Ferguson — UK Centre for Ecology and Hydrology: https://www.ceh.ac.uk/ LinkedIn: / shona-ferguson ——————————————— AI in Production is Jumping Rivers' conference for data scientists, ML engineers and AI practitioners building and deploying AI & ML in the real world. The inaugural event was held at The Catalyst, Newcastle Helix, on 4–5 June 2026. 🎟 Event & full programme: https://ai-in-production-2026.jumping... ▶ More talks from the conference: / @jumping-rivers 📑 Full playlist — every AIP 2026 talk: • AI In Production 2026 🌊 Organised by Jumping Rivers: https://www.jumpingrivers.com/ With thanks to our sponsors — Databricks, Posit, the Royal Statistical Society, Chapman & Hall/CRC, and the National Innovation Centre for Data. Community partner: DevITJobs.uk. #AIinProduction #AIP2026 #MachineLearning #LLMs #AIEngineering

Andrej Karpathy: From Vibe Coding to Agentic Engineering w/ Stephanie Zhan

Quantum Physicists Just Broke The No-Cloning Theorem

The World's Most Important Machine

Agentic AI Frameworks Explained: Workflows, Multi-Agent, & Production

8) Programming with LLMs in R & Python | Jumping Rivers Webinar Series

Something is jamming GPS over Europe. Here's what we found

Async Agents in Production: Failure Modes and Fixes — Seb Ringrose | AI in Production 2026

Stop Prompting Claude. Use Karpathy's Method Instead.

But what is the Fourier Transform? A visual introduction.

Why AI Agents are either the best or worst thing we’ve ever built

Has an AI discovered new maths?

🚨 If UK Police Stop You and Demand Your Name — Say THIS (You Don't Always Have to Give It)

Yann LeCun: World Models: Enabling the next AI revolution

But what is quantum computing? (Grover's Algorithm)

This New Google Format Gives Your AI Agent a Second Brain

Billionaire's WARNING: I'm SELLING. The Crash Is Already Here!

The best AI agents are simpler than you think

Don't learn AI Agents without Learning these Fundamentals

Full AI Prompting Course with Andrew Ng

