generative-AI-coding-course

Generative AI for writing (research) software

This is the course material for the course “Generative AI for writing (research) software”.

Course content: Generative AI is emerging as a major creative force that supports humans in content creation. Specifically trained models combined with tools and workflows can support software developers with their software projects. Use of these AI agents can lead to time savings and a shift in what aspects of generating software are more important on a day-to-day basis. In this course, we will learn how to set up and use AI agents in software development projects. Best practices in using such tools, as well as recommendations how to use them efficiently and safely will be introduced. The slides can be found here.

For the short version of the course, the slides can be found here.

Learning goals:

• Use AI-assistance in a coding environment
• Know about the limitations of these tools
• Know how to adapt AI agents
• Know how to provide additional tools to AI agents
• Be aware of legal and ethical implications
• Be aware of privacy and security concerns in the use of such tools

Prerequisites: Basic Python knowledge is required. Participants need a laptop/PC with Visual Studio Code installed and a working Python environment. Participants need to have access to GitHub Copilot (either through free trial, individual license such as GitHub student (free), or other form of license).

How to get access to GitHub Copilot

Before the course, you need to sign up for a GitHub Copilot license. There are several options:

• Sign up for a free trial: You can get limited free access to GitHub Copilot
• Sign up as GitHub student: If you are currently a student/PhD student, you can get access through signing up for GitHub student developer pack
• Sign up as GitHub teacher: If you are currently a Postdoc or similar, you can get access through signing up for GitHub teacher
• Sign up for a paid individual license
• Access GitHub Copilot through your Copilot Business or Copilot Enterprise subscription

Course Content

Intro to generative AI for coding tasks

• Generative AI for software development
• Providers versus tools versus models
• AI agents versus AI models

Using an IDE with an AI agent

• Basic set-up and capabilities of GitHub Copilot with VSCode
• Example 1: Basic usage
• Example 2: More advanced code with numpy
• Using AI agents: Good practices
• Software validation and verification
• Example 2: User story

Adapting AI agents

• The agent core identity
• System prompts
• AGENTS.md
• Agent modes
• Agent context
• Agent loop
• slash commands
• Example 3: English quotes dataset

Demonstrations

• Demo with Cursor
• Demo with Claude CLI
• Demo with Antigravity

Agent tools

• SKILLS.md
• Example 4: English quotes skill
• Model-Context protocol (MCP)
• Example 5: Using the knowledgebase MCP
• Retrieval-augmented Generation (RAG)

Beyond the technical

• Security aspects
• Legal aspects
• Ethical aspects
• Engineeringification
• The future of research?

Summary

• Reflecting on the learning goals

This site is open source. Improve this page.