Mental Health RAG (Retrieval-Augmented Generation) Assistant

This project is a domain-specific Retrieval-Augmented Generation (RAG) system built to assist with mental health queries by leveraging a structured knowledge base derived from the DSM-5 (Diagnostic and Statistical Manual of Mental Disorders). The assistant is capable of providing grounded and contextually relevant responses by combining prompt engineering with vector-based semantic retrieval.

Key Features:

Domain Knowledge Ingestion: Parsed and preprocessed content from the DSM-5 textbook to serve as the foundational knowledge source, allowing the assistant to emulate expert-level understanding in clinical psychology and psychiatry.

RAG Architecture: Implemented a hybrid retrieval-generation system using embeddings and vector search to fetch relevant context from the DSM-5 before passing it to an LLM for final response generation.

Prompt Engineering: Crafted effective and structured prompts to guide the LLM in producing accurate, ethical, and relevant mental health advice grounded in the DSM-5 context.

Vector Store Integration: Utilized embedding models to convert text chunks into high-dimensional vectors and stored them in a vector database for efficient similarity-based retrieval.