Clinical Trial RAG Assistant — Retrieval-Augmented Summarization of Medical Studies

A Retrieval-Augmented Generation (RAG) system for querying and summarizing real-world clinical trials in clear, patient-friendly language.

This project ingests data from ClinicalTrials.gov and combines:

Semantic retrieval using SentenceTransformers and ChromaDB

Local open-source LLMs (Gemma / Phi) for grounded text generation

A lightweight Gradio web interface for interactive exploration

Users can ask natural-language questions (e.g. “diabetes trials in Europe”) and receive concise summaries synthesized from multiple relevant studies, with filters for conditions and adjustable retrieval depth.

Designed as an applied NLP + healthcare ML system, emphasizing:

End-to-end data pipelines (API ingestion → parsing → embeddings → inference)

Practical RAG architecture

Responsible AI usage in medical contexts