Project : CustomChatBots

[tareeb]

Project Name

CustomChatBots

Description

We developed different pipelines to try out both simple and advanced RAG methods. The simple pipeline uses Grok-LLama3, Hugging Face’s embedding model, and ChromaDB for storing vectors. The advanced pipeline includes newer RAG techniques like multi-query retrievers and parent-child chunking, and it uses Neo4j and Azure OpenAI.

Our main focus was to make RAG easy to use and easy to work with. We wanted to move RAG from Colab Notebooks to a production environment. So we designed a client-server architecture so it can be used as a product by real users. Just like Langchain connects different parts, we provide a ready-made setup and frontend components to create RAG-based chatbots that you can easily add to your websites. Our backend is also flexible, allowing you to update and replace different components easily.

Summary :
Using CustomChatBots, users can effortlessly create chatbots that follow custom instructions and utilize their custom data as a knowledgebase. Updating the data and instructions is simple and can be done anytime with just a few clicks. After creating a chatbot, user receive a public URL that they can integrate into their own user interfaces. Alternatively, they can also use our ready-made React components in their projects easily.

The more detail about the project description and the motivation, and how it work can be in readme.md of github repository.

• Details: CustomChatBot
• Detail Screenshots are given to demonstrate how easily chatbots can be created, and instructions and data can be kept up to data with simple clicks

Technology & Languages

• JavaScript
• Java
• .NET
• Python
• AI Studio
• AI Search
• PostgreSQL
• Cosmos DB
• Azure SQL

Project Repository URL

https://github.com/tareeb/CustomChatBot-backend

Deployed Endpoint URL

No response

Project Video

https://youtu.be/c2GdtBP1FmI

Team Members

@hasnatahmed331

[multispark]

Hello @tareeb, @hasnatahmed331, thank you for participating in RAG Hack!

The team is working hard to distribute badges. Please have each team member fill out this form:
aka.ms/raghack/badge-dist

Thank you!