notes.txt

Team Air
Level 1: Basic RAG
- Setup Knowledge Base with 5 papers
- Create an index database
- test the chatbot using a large language model

Level 2: Custom UI
- Build a custom interface
- Optimize parameters to get better results
- Expand Literature collection (thousands of papers)

Level 3: Enhanced Agent
- Be creative and provide custom instructions for a supportive agent,
e.g. personal tutor, text creation=)
- So add modes to the LLM class that uses different system prompts?

S3 Bucket for data
Amazon Bedrock for ML API


Set the team name air on everything you create
1. Set up storage for your papers with S3
2. Build knowledge base in bedrock
3. Use AWS OpenSearch service to create an index database
4. Probably use cohere embedding model and claude 3.5 sonnet as the text model
5. Push website to another s3 instance and use AWS cloud front
6. Use AWS Lambda to run the chatbot