AI-Powered Knowledge Assistant with RAG by Marco EAI-Powered Knowledge Assistant with RAG by Marco E

AI-Powered Knowledge Assistant with RAG

Portfolio project
Marco E
Developed an internal AI-powered knowledge assistant for Boehringer Ingelheim to automate recurring helpdesk inquiries.
Built a Retrieval-Augmented Generation (RAG) pipeline that processes PDFs, Excel files, and other knowledge sources, using Pinecone for vector search and relevant information retrieval.
Implemented intelligent request routing, structured LLM responses, and email escalation for questions that couldn't be answered automatically. The solution included a React frontend and was evaluated against real helpdesk questions with reference answers.
Technologies: Python, RAG, LLMs, Pinecone, React, Document Processing

0 likes · 0 views · Nov 10, 2026

Tags