This masterclass provides an intensive deep dive into new developments in LLMs that introduce a new paradigm involving reasoning and agentic workflows.
The first generation of language models are sometimes characterised as “stochastic parrots”, since they generate outputs based on what word is most likely to come next, given all of their training data.
However, new foundation model paradigms are emerging that devote more computation time to generating their output. These models, termed reasoning models, can provide more in-depth answers to questions. Another exciting development is that of agentic workflows, which can involve multiple LLM-based agents that are assigned distinct roles. Again, by devoting substantial computing resources to answering a single question, they can generate much richer results in response to an initial request.
This half-day workshop is designed for professional people who already are familiar with using language models. You will begin by exploring the topic of co-creating with generative AI tools, and then learn in depth about new reasoning modes and agentic AI. Finally, you will put your learning into practice with a hands-on session.
Participants must bring a laptop (fully charged) and already have an account (free-tier will be enough) set up on one or more advanced LLM products that include reasoning, such as ChatGPT, DeepSeek, or Claude.
