Inspiration
I was experiencing a lot of pain in a small startup environment because it was hard to keep track of what other people are working on. Then, when I get assigned to those projects, it take a long time ramp back up because my knowledge becomes outdated, and I have to look into the codebase again, ask questions to others to get back to the playing field.
So, if there is a way to keep track of projects or issues I was involved or am interested in more frequently, it would save me a lot of time in the future, and potentially give help to people that need.
What it does
- Issue content is embedded and saved to vector database whenever new issue is created. (simplified in the system design diagram)
- When the application is activated (by pressing the button), backend find relevant issues from embeddings stored in the vector database.
- Similar issues are returned and passed to the Jira application.
- Jira application creates a comment with the relevant issue url and tags the issue assignees.
How we built it
The system diagram is posted as image.
We used forge to build the UI on Jira. We built a backend server using Lambda (deployed with Terraform) that embeds the issue content using OpenAI embedding API. We save the embedding to Pinecone (vector db) and fetches similar issues from the past. Then, we create a comment stating those relevant issues with the issue link and tags of the issue assignees.
Challenges we ran into
Getting used to Forge was the biggest challenge.
What's next for StayConnected
- Cross-platform application to take activities on Confluence and BitBucket into account
- Enhanced embedding strategies to improve quality of connections
- Improved user experience of explaining and displaying connections with further automation
Built With
- amazon-web-services
- forge
- openai
- pinecone
- terraform
Log in or sign up for Devpost to join the conversation.