Understanding and Building Large Language Models 6 ECTS Spring 2026 (IDA)

---

Course Examiner and Lecturer: Fredrik Heintz, Linköping University

The aim of the course is to explain the methods and techniques used by large generative AI models such as large language models work and explore how to build them. The focus is on the technical aspects such as methods and techniques. The course is thus more about machine learning than about natural language processing.

Full course plan 2026.

---

Examination

The examination consists of LAB1 with four lab assignments. You are welcome to work alone or in pairs.

LAB1: Develop an LLM from scratch and conduct at least two experiments related to the lectures and write a short report on it. The lab is expected to have four parts:

• Lab 1: Develop a simple data pre-processing pipeline.
• Lab 2: Pre-train a GPT-style LLM.
• Lab 3: Fine-tune the LLM.
• Lab 4: Evaluate the LLM

---

Schedule 2026

The lectures will normally be both in person at IDA (see room below) and online on Zoom.

• 2/3 15-17 Alan Turing, Introduction to NLP and LLMs, slides
• 9/3 15-17 online only, Basics, slides
• 16/3 15-17 Alan Turing, Data curation and processing, old slides
• 23/3 15-17 online only, Pre-training and scaling, old slides
• 30/3 15-17 Alan Turing, Fine-tuning, aligning and distillation, old slides
• 20/4 15-17 Alan Turing, Inference, in context learning and retrieval augmented generation, old slides
• 27/4 15-17 Alan Turing, Benchmarking and evaluation, old slides
• 4/5 15-17 Alan Turing, Building LLMs in practice part 1 with TBA, old slides
• 11/5 15-17 Alan Turing, Building LLMs in practice part 2 with TBA, old slides
• 18/5 15-17 Alan Turing, Multi-Modal LLMs, old slides

---