LangChain Hackathon Summary

A Retrieval-augmented generation (RAG) that will answer users about Coursera courses, this RAG was made using langcahin, Gemini pro and pinecone as a vector database. we got an inspiration to make this RAG because we as students always look for a way to learn, and an AI system that answer our questions about Coursera courses would be beneficial for us and other students who look for courses. We have utilized Pandas to do some cleaning before inserting the courses data inside the vector database, and made sure that the data was clean and ready in a format that would be suitable to be queried by a vector database

Data Piece

the main idea is to help extracting features and further data analysis for any given structured dataset for example when a data analyst has the data in hands and needs more in depth analysis, the LLM helps him by giving him further instructions, given the data and a short dataset description, the LLM generates questions and ideas on how to further analyse the data to make a more performing model for whatever the use case of the dataset, time series, labeled and unlabeled, or semi labeled. for that, the data analyst can read the instructions generated by the LLM and then apply them in order to help him solve the problem.

Sinx