Skip to content

Schedule

Here is the schedule for the Winter School. Tuesday - Thursday will take place entirely at Jesus College. On Friday morning, there will be one session in Elena Hall at Jesus College, and the afternoon will be at the William Gates Building on the Cambridge West Site.

Tuesday January 13

Time Title Description
15.00 - 16.00 Arrival and check in Settle in to your accommodation at Jesus College
16.00 - 18.00 Welcome reception Get to know your fellow Winter School AI practitioners over drinks and a buffet dinner!

Wednesday January 14

Time Title Description
09.30 - 10.30 Intro to LLMs In the morning, we walk through some of the theoretical concepts that underpin modern Large Language Models. We will cover: tokenization, embeddings, the attention mechanism, and sampling and pretraining.
10.30 - 11.00 Break
11.00 - 12.30 Intro to LLMs Session continues
12.30 - 13.30 Lunch
13.30 - 15.00 Hands on LLMs Grab an API Key and play around with some models. We cover the OpenAI API, basic prompting, and tool use.
15.00 - 15.30 Break
15.30 - 17.00 Hands on LLMs Session continues
18.00 - 19.00 Dinner Buffet dinner and drinks

Thursday January 15

Time Title Description
09.30 - 10.30 Research Software Engineering This workshop offers a high-level overview of core software engineering practices that enable reproducible, trustworthy research software, with particular relevance to modern AI research and development. After introducing the Oxford RSE group and outlining our work with the Schmidt AI in Science Fellows, we will present a condensed version of some of the training we provide to the Fellows. This will include the disciplined use of version control to support traceability and collaborative development using platforms such as GitHub. We will also discuss the use of testing and continuous integration to enforce standards and support reproducibility. Together, these techniques form essential infrastructure for sustainable software development and provide a foundation for high-quality research outputs.
10.30 - 11.00 Break
11.00 - 12.30 Research Software Engineering Session continues
12.30 - 13.30 Lunch
13.30 - 15.00 Responsible AI for Coding This workshop guides attendees on integrating AI into software engineering workflows effectively and ethically. Moving beyond basic code generation, the session explores LLMs—from web interfaces and IDE extensions to agentic CLI tools. Participants learn to balance AI's opportunities for rapid prototyping against risks like security vulnerabilities, hallucinations, and "sycophantic" bad code. While AI serves as an "enthusiastic but inexperienced intern," ultimate responsibility for code correctness, security, and maintainability remains with the developer.
15.00 - 15.30 Break
15.30 - 17.00 Responsible AI for Coding Session continues
18.00 - 22.00 Dinner Dinner at Jesus College

Friday January 16

Time Title Description
08.30 - 09.30 Checkout Checkout out of accommodation. You will not be able to gain access to your rooms after this time.
09.30 - 10.30 An Introduction to GPUs Have you ever wondered about the technology that made the explosion in deep learning possible? In this session, experts from NVIDIA will walk through some of the key concepts behind GPUs.
10.30 - 11.00 Break
11.00 - 12.00 An Introduction to GPUs Session continues
12.00 - 12.30 Lunch After lunch, we will move from Jesus College to the West Cambridge Site. You will not be able to access Jesus after this time, and your luggage will be stored safely in West Cambridge.
12.30 - 13.00 Transport to West Site For the final part of the day, we need to travel over to the William Gates Building at the West Cambridge Site.
13.00 - 15.00 Introduction to HPC In this session, Research Computing Services (RCS) will introduce high performance computing (HPC) at Cambridge. This will include an overview of the West Cambridge Data Centre, and a practical introduction to accessing the Dawn Supercomputer.
15.00 - 15.30 Break
15.30 - 17.00 Running AI Applications on Dawn After a tour of the West Cambridge Data Centre, we will run a diffusion model on the Dawn Supercomputer.