Close the GenAI Intent <> Impact Gap

Beyond hype, beyond skepticism. This is not just about learning how to use a specific tool. This is about understanding the new normal.
A masked superhero floats in the area, with a star on his chest and a fluttering cape, above the word "Expectation". Next to him, is a child wearing the same star and cape, but half the height, and standing on the ground, above the word "Reality".

Our Hot Take

There’s so much hype, so much value, so many benefits, and so many dangers. If you want your software engineering team to take advantage swiftly while mitigating the risks, the most powerful thing you can give them is a deep understanding of the benefits and risks of GenAI. Then they can use it rationally, evaluating and appropriately using new AI tooling as it is released.

Learning, with impact

Learning Outcomes: After this series of hands-on, interactive micro-workshops, engineering teams will be able to:
  • Understand what generative AI is and how it applies to their work
  • Identify good and risky use cases for different forms of AI
  • Demonstrate a deep understanding of how LLMs work and why they work the way they do
  • Develop effective strategies, competence and confidence experimenting with, and incorporating new AI tools
  • Appreciate the role of training and RLHF in the evolution of an LLM
  • Solve a well-defined problem using LLMs to generate code
  • Confidently and effectively debug code using LLMs

Session Outline

Session 1: Intro to Generative AI (90 mins)

This hands-on micro-workshop will introduce AI and its different forms, and the power of LLMs. It will also highlight some of the potential dangers of LLM usage.

  • What machine learning (ML) and artificial intelligence (AI) are, at a high level.
  • The various forms of AI, and ways in which they are used.
  • The strengths and weaknesses of generative AI tools, including LLMs.
  • How AI can be applied to work.
  • How to identify good and risky use cases for different forms of AI.
  • How to evaluate the output of LLMs using different metrics and why it’s important to do so.
Session 2: Under the skin of LLMs (90 mins)

This micro-workshop demonstrates the fundamental building blocks of Large Language Models (LLMs) - tokenizers, transformers, attention. These structures are tied to core ideas of prompt engineering - writing better prompts and interpreting the behaviour of LLMs.

  • The key components of an LLM and the transformer model, including tokens, attention.
  • How LLMs are trained, and how this influences their strengths and weaknesses.
  • How to prompt LLMs effectively - including setting the temperature, top-p and top-k sampling, self-consistency, and multi-shot prompting.
Session 3: Coding with LLMs (90 mins)

A hands-on micro-workshop focused on using LLM-based tooling effectively in the software development life cycle.

  • Different ways to use LLMs to solve problems in software engineering.
  • How to build conversationally, how to generate new code and generate changes to code.
  • How to debug code using LLMs.
Session 4: Deep Dive into LLM Tools (90 mins)

A micro-workshop that will dive into some specific LLM tools for writing software, and how to use them effectively.

  • How to use a range of different tools for software development.
  • How agentic AI software development systems work.
  • How to use LLM-centric IDEs to make changes to multiple files at once.
Session 5: Integrating with LLMs (90 mins)

This micro-workshop focuses on building software that integrates with LLMs.

  • Key considerations when integrating with LLMs (e.g. sanitisation, escaping, designing for human oversight).
  • How to build prompt templates.
  • Identifying and mitigating the risks when building software on top of LLMs.
Session 6: CustoMISING AI Systems (90 mins)

This micro-workshop explores how to build your own AI systems using LLMs.

  • How to fine-tune a foundational LLM.
  • How to design and implement RAG systems.
  • Grow familiarity with some off-the-shelf tools for RAG and fine-tuning.

Coming in Q3 2025

  • GenAI Excellence for non-Engineering Teams (6 modules)
  • Building AI for Engineering Teams

This is the future of learning