Session Plan

The idea behind this session is to teach how AI can be used in the day-to-day development workflows of a developer.

Start by giving a brief overview of a usual workflow, and then dive into specific parts of it and how AI can help there, with specific prompts and exercises to practice. The suggestion is that mentors talk through specific use cases, and demo the prompts in the session. The exercises are designed for the trainees to practice themselves on their own machines.

AI in the development workflow

Workflow overview

Personal development - Reading, learning, practicing new software development skills
Problem definition - Understanding what we're trying to solve
Planning - Deciding what to build
Design - Deciding how it should work and look
Implementation - Building the software
Testing - Ensuring it all works correctly
Code review - Getting feedback and finalising the code
Deployment - Shipping it to users
Monitoring - Making sure things keep working as intended

After the overview, go through some specific phases to highlight key ways AI can support them. Of course there are many possiblilities in every phase (and we should encourage trainees to explore them themselves!), but the below are mostly aimed to inspire and give some concrete examples in a few of the phases that trainees can start using today.

Personal development - Learning

Outside of HYF or work, spending time learning and practicing new development skills is important. This could be learning new programming languages, frameworks, tools or concepts.

Here's some examples:

1. "Explaining by comparison" prompt

Learn quickly by mapping new knowledge to things you already understand.

I know how to do X in <this language I already know>:

<paste code snippet here>

Show me how to write the same functionality in <new language> and explain:
1. What concepts are the same
2. What's different and why
3. What would surprise a developer coming from <language I already know>

2. The "Code Tutor" prompt

Learn a new language/framework/tool through practice, and getting feedback to improve.

I'm learning how to do X in <language/framework/tool>. Here is my current solution:

<paste code snippet here>

Review this as if I were a junior developer:
1. Point out any issues, bugs or bad practices
2. Suggest better practices or patterns
3. Explain what each change means and why it matters
4. Create a mini summary of what I should study next, based on my weaknesses and mistakes

3. The "Learning plan" prompt

If you have a big topic ahead of you, get help to create a learning path to help you stay on track and avoid becoming overwhelemed.

I want to learn X as a developer who already knows Y and Z. Create a 4 week self-study plan focussed on hands-on coding practice.

For each week, include:
1. Key concepts to learn
2. One small practical project
3. Common mistakes and misunderstandings beginners make
4. How this new thing differs from what I already know

Exericse 1

Use each of the three prompts above using your own personal learning goals, and evaluate what you like and don't like about the output of the AI. Bonus: Try tweaking a prompt template to shape the output more to your own personal needs.

Top tip If AI writes code you can’t explain:

Slow down
Ask why
Rewrite it yourself until you understand

That is the moment where learning happens, so don't miss out by skipping ahead!

Design - Code explanation

While preparing ideas for a particular solution in the Design phase, you'll likely come across the challenge of needing to understand existing code.

TODO add some prompting examples here, as well as good follow up prompts

Exercise 2

Use AI to help explain the code in [Exercise 2](TODO create and insert link here). Ask follow up questions until you have a good grasp of what every function and line of code is achieving.

TODO write the exercise code, inspired by exercise

Implementation - Learning new approaches

While writing code, you may come across a roadblock where you're not entirely sure how to implement something.