Microsoft Developers AI Learning Hackathon Phase 1

Azure OpenAI Python Developer Guide Proof of Completion
Azure Cosmos DB Developer Cloud Skills Challenge Part 1 Proof of Completion
Azure Cosmos DB Developer Cloud Skills Challenge Part 2 Proof of Completion

Inspiration

I joined this hackathon because I was very intrigued by how OpenAI integrates with an enterprise-scale cloud platform like Azure. I have also used Azure before as an IT administrator, and I was eager to use it in a development context, as well.

What it does

This submission is a confirmation that I have completed the requirements for Phase 1, which includes the two-part Azure Cosmos DB Developer Cloud Skills Challenge and the Step-by-Step Developer Guide to build a custom AI copilot.

How I built it

I read over all the content in the challenges and guides and applied it in the included labs. I tested and made sure that the backend and frontend components worked at the end of the labs.

Challenges I ran into

A platform like Azure is very strict with permissions and quotas, so it was challenging and frustrating to be hit with blocks due to these restrictions. Over time, I became more skilled with identifying these kinds of issues, so they became increasingly less cumbersome.

I also experienced the difficulty of handling errors generated by automated deployments with the bicep template, which was very challenging to debug at first. Once I got it working, however, it was impressive how it could build up a whole environment in a matter of moments!

Accomplishments that I'm proud of

I was proud of being able to complete the Phase 1 challenges from end-to-end, taking in as much knowledge and expertise as I could in the process. Some roadblocks were so challenging that I was tempted to quit, so I am proud that I stuck through and made it to the end.

What I learned

I learned a lot about using and integrating generative AI into a scalable project using Azure, how to leverage RAG and vector-based solutions like CosmosDB vcore for querying AI-powered solutions, and string sequences of calls with Langchain.