Machine learning engineer for AI chatbot development and document retrieval system

Job Description (AI Chatbot RAG system) We are looking for a Machine Learning Engineer to build a RAG-based chatbot system using a local LLM that can answer questions from a collection of PDF documents. Project Overview We have a dataset consisting of multiple PDF files (mixed structure: text-heavy documents, notices, and tabular information inside PDFs). The goal is to build a system where users can ask natural language questions and get accurate answers grounded strictly in the content of these PDFs. The system must: Extract and process text from PDFs Chunk and index the content for retrieval Use embeddings + vector search for relevant context retrieval Use a local LLM (not API-based) to generate answers Ensure answers are strictly grounded in retrieved content (minimize hallucination) Key Responsibilities Design and implement a RAG pipeline for PDF documents Build a robust PDF parsing and chunking system Create embedding + vector database pipeline (FAISS, Chroma, etc.) Integrate a local LLM for inference (Ollama, vLLM, Transformers, etc.) Optimize retrieval quality and context selection Ensure system is stable and production-ready (not just a demo script) Required Skills Strong experience with RAG systems Experience with PDF parsing / document ingestion Knowledge of embeddings and vector search Experience with local LLMs (Ollama / vLLM / HuggingFace Transformers) Strong Python skills Nice to Have Experience with hybrid retrieval (BM25 + vector search) Experience handling noisy PDFs (OCR, scanned documents) Understanding of chunking strategies and retrieval evaluation Deliverables End-to-end working RAG chatbot PDF ingestion + preprocessing pipeline Vector index setup Local LLM integration Simple documentation explaining architecture How We Work We prefer engineers who think in terms of system design, not just libraries. Important Note This is not a basic chatbot or prompt engineering task. We are specifically looking for someone with real experience building RAG-based document intelligence systems.