Skip to content

An RAG-based conversational AI app helping students navigate the complexities of higher education.

Notifications You must be signed in to change notification settings

AI-Student-Advisor/AI-Student-Advisor-Backend

Repository files navigation

AI Student Advisor - AI Component

The AI component of the AI Student Advisor is responsible for providing the chat engine with the ability to answer questions based on the data and remember conversation history.

Primary Stages

  • Data Loading: Ingest data, generate embeddings using LLM model, and store in a hosted vector database
  • Context Retrieval: Retieve relevant context from vector database based on user query
  • Response Generation: Generate response based on user query, retrieved context and previous conversation history

Components Required for Chat Agent

  • Retrieve & Load Data: Use data loaders to create a retrieval tool for the chat engine
  • Embedding Generation: Generate embeddings for the source data using the configured embedding model
  • Chat Agent Instantiation: Create a chat agent that can answer questions based on the data & remember conversation history

Chat Agent Architecture

User Query Handling Workflow

Chat Agent User Query Handling

API

User Query Handling

User Query Handling