Working With LLMs: course outline

A hands-on, self-paced course in getting reliable work out of AI models, knowing exactly when they fail, and never handing over your own thinking.

1. Module 0. Start here

   15 min

   How the course works, and a quick look at where you are now

   1. 0.1 How this course works Free preview5 min
   2. 0.2 Where are you now? Free preview10 min

2. Module 1. How LLMs actually work

   89 min

   The one-sentence mental model that predicts their behaviour

   1. 1.1 What an LLM is and isn't Free preview10 min
   2. 1.2 Tokens and the generation loop 12 min
   3. 1.3 How models are trained 10 min
   4. 1.4 Sampling, cutoffs and reasoning models 12 min
   5. 1.5 Six facts that change how you use it 8 min
   6. 1.6 Lab 1: Break it, then explain it Lab 25 min
   7. 1.7 Prove it 12 min

3. Module 2. Prompting fundamentals

   77 min

   Writing the brief you would give a capable stranger

   1. 2.1 The GCEF brief 12 min
   2. 2.2 Spot the missing element 8 min
   3. 2.3 Habits of skilled users 12 min
   4. 2.4 The document you never pasted 8 min
   5. 2.5 Lab 2: The brief Lab 25 min
   6. 2.6 Prove it 12 min

4. Module 3. Advanced techniques

   82 min

   Matching the technique to the shape of the task

   1. 3.1 Matching technique to task 12 min
   2. 3.2 Watch a technique change the answer 10 min
   3. 3.3 The multi-prompt workflow 12 min
   4. 3.4 Careful reasoning, false premise 8 min
   5. 3.5 Lab 3: Technique selection Lab 30 min
   6. 3.6 Prove it 10 min

5. Module 4. Retrieval, tools and agents

   94 min

   Letting the model read, act and work in steps, safely

   1. 4.1 Retrieval 12 min
   2. 4.2 Lab 4: Tiny RAG, no code Lab 30 min
   3. 4.3 Tools 10 min
   4. 4.4 Agents and the ReAct loop 12 min
   5. 4.5 Prompt injection and safe use 12 min
   6. 4.6 Which addition fixes it? 8 min
   7. 4.7 Prove it 10 min
   8. 4.8 Optional: the same lab in code 10 min

6. Module 5. Verification

   114 min

   Knowing what to check, and measuring reliability

   1. 5.1 Why models fabricate 10 min
   2. 5.2 The trust ladder 10 min
   3. 5.3 Verification techniques 14 min
   4. 5.4 Lab 5: The audit Lab 30 min
   5. 5.5 Building a small evaluation 10 min
   6. 5.6 Lab 5b: Your first eval Lab 30 min
   7. 5.7 Prove it 10 min

7. Module 6. Ethics and responsible use

   69 min

   Deciding whether to use the model at all

   1. 6.1 The delegation decision 12 min
   2. 6.2 Integrity and the disclosure line 10 min
   3. 6.3 Privacy and bias 12 min
   4. 6.4 Lab 6: The policy you'd sign Lab 25 min
   5. 6.5 Prove it 10 min

8. Module 7. LLMs in your field

   157 min

   Building and testing your own workflow

   1. 7.1 The five questions 10 min
   2. 7.2 Workflows in your field 12 min
   3. 7.3 Lab 7: Design your workflow Lab 30 min
   4. 7.4 Project clinic: three runs Lab 45 min
   5. 7.5 Capstone: explain it to an outsider 40 min
   6. 7.6 Finish: re-test, reflection, certificate 20 min