Bedrock + Wing Workshop

Building a Serverless Workflow, Powered by AI

This repo is a step-by-step guide for an online workshop (Eventbrite).

This workshop is available on youtube

Any issues? Questions? Let's talk!

Please join Winglang's Slack; we have a dedicated channel for this workshop #workshop-bedrock.

During the session, you'll learn the following

• Build a GitHub App and Deploy it to AWS (using terraform)
• Introduction to Amazon Bedrock
• Using Amazon Bedrock resources from Wing
• Creating a GitHub App from scrath, developing it locally

Some General Resources

• Learn about Inflights - Preflight and Inflight are Winglang's most important core concepts.
• Cheat Sheet - Let's learn Wing in 5 minutes, shall we?
• Language Reference - A complete language reference for Winglang.
• Winglibs - Wing Trusted Library Ecosystem
• Playground - Write and test Wing code online.

The Application

A GitHub Application (AKA GitHub bot) that listens to any incoming pull requests and corrects spelling, grammar, & punctuation for any *.md files changed during the pull request.

See Demo

Workshop Sessions

1. Setup & Prerequisites - Tools, setup, and getting access to AWS Bedrock Foundation Models (link).
2. Why Winglang - Problems with serverless, and the Wing approach (pdf).
3. Introduction to Bedrock - A short introduction to Amazon Bedrock service by Arik Porat, Solutions Architect at AWS.
4. Using Bedrock in Wing - Installing & using the winglibs/bedrock module (link).
5. Create a GitHub App - Step by step guide to creating a GitHub app (link).
6. Wing's GitHub Module - Using winglibs/github instead to make an uppercase bot (link).
7. Spell Checking with Bedrock - Modifying the files based on winglibs/bedrock (link).
8. Deploy to AWS - Deploying the application to AWS (link).

---

What's Next?

In the next session, we will be using LangChain with AWS Bedrock to embed documents into Momento Vector Index.