WUZETIAN

Inspiration

I was inspired by one of my previous experience of building a health agent using LLM model API's. By preparing an organized prior knowledge and managing chat history, chatbots could demonstrate satisfactory performance.

What I Learned

Using Integration with AWS Bedrock: I gained experience of working with Amazon Bedrock, managing the account, getting access to and comparing among cutting-edge LLM models. Data Handling: I gained experience on managing large datasets, especially breaking down information into manageable chunks to avoid token limits.

How I Built It

The project pulls information from CSV files and text files containing agents, weapons, game rules, and players. This data is chunked and fed into the model's memory to prepare it for upcoming questions.

Challenges I Faced

Token Limits: One of the biggest technical challenges was handling the model's token limit, which required me to split the data while preserving the information logic. New to Amazon AWS: It was complicated at first to figure out setting both the AWS account itself, creating roles for controllable token access, and registering the credentials locally.

Built With

• python