Fall 2026, running from September 1 through December 8, 2026. Classes meet twice a week, on Tuesdays and Fridays. Each week corresponds to a chapter (or pair of chapters) of the course textbook, Agentic Large Language Models. Winter break follows the December 8 session.
Exact times and rooms: This page is a topic and deadline guide, not the authoritative timetable. For the definitive class times, rooms, and any last-minute changes, check Brightspace and rooster.universiteitleiden.nl.
The chapter groupings for weeks 5–7 (Reasoning & Agents) reflect how the teaching team compresses Parts II and III of the book into three lectures; the exact week-by-week split within that block may shift slightly as the course is finalized.
Click a group below to see its weekly topics and assignment dates. For the book chapters behind each topic, see the Syllabus.
Sep 1
Introduction
What is an agentic LLM? The reason/act/interact taxonomy and the "harness" concept.
Sep 8
Language Modeling I
Tokenization, embeddings, attention, the transformer architecture. Assignment 1 starts (Building LLMs from Scratch).
Sep 15
Language Modeling II
The training pipeline — data, pretraining, finetuning, preference tuning, inference, benchmarks.