Responsible AI refers to the set of defined principles and applied practices that help us ensure AI remains safe, reliable, transparent and fair, in ways that live up to users' trust and expectations. Traditionally, AI focused on predictive models making responsible AI a key part of the data scientist toolkit for model debugging and decision-making. With the growth of large language models, the focus now moves to developers and generative AI applications - and integrating responsible AI tools and techniques into end-to-end developer workflows (LLM Ops).
My goal with this project is to document my learning journey into Responsible AI for Generative AI Application Development and capture those insights in the form of easy-to-use "recipes" or hands-on labs that I (and other learners) can revisit for code snippets or practical tips on demand. I have 3 main learning objectives:
- Understand core terminology and concepts for Responsible AI
- Get hands-on experience with core tools for assessing Responsible AI
- Learn to integrate and use Responsible AI tools in LLM Ops workflows
![WARNING] This is primarily intended as a learning resource and sandbox for exploration and is not intended to be used in production. It is my learning sandbox so expect breaking changes. Use issues for feedback.
To get the most from this resource, you will need:
- A GitHub account
- An Azure account (for AI integrations)
- Familiarity with Python and Jupyter Notebooks
- Familiarity with Visual Studio Code Extensions
I'll use this section to document any dependencies required to explore the hands-on segments on your own. This repository is instrumented with a devcontainer.json configuration that provides a prebuilt development environment for use. Just fork the repo (to your personal profile) and launch GitHub Codespaces for a quick start.
Check back for an updated Roadmap on what we will cover. Visit the website to see blog posts related to Responsible AI usage, tools and announcements - and for updates on this project.