Created the website using React and integrated with Auth0
Using Langchain to convert PDFs to text documents and then break it down into more manageable chunks to read into Pinecone
Using the Pinecone Node.js packages to query data from the PDFs based on user prompts
Using the OpenAI APIs to convert the user prompt to vectors to query Pinecone.
Wrote the backend in Node.js deployed to AWS Lambda and API Gateway, using DynamoDB to store non-vectored data. Also using S3 to store the files that are uploaded.
