I will build a custom ai chatbot for your documents using rag and fastapi

M
mohamedhany575
M
mohamedhany575
Mohamed Hany

About this gig

Most basic RAG chatbots fail in production because simple vector search misses exact product codes, clauses, or follow-up context. I build production-ready FastAPI + Qdrant retrieval pipelines engineered for real data.

What I Build For You:

  • Hybrid Search (Dense + BM25): Captures both conceptual meaning and exact part numbers, SKUs, names, and clauses that standard bots miss.
  • Two-Stage Reranking: Every chunk is re-scored via a Cross-Encoder reranker before reaching the LLM, cutting hallucinations and returning exact source citations.
  • Follow-Up Memory: Automatically rewrites multi-turn follow-up questions into standalone search queries.
  • 100% Flexible LLM Backend: Works with OpenAI, Groq, OpenRouter, or private local models (Ollama, LM Studio, vLLM) so you control your data and API costs.

What You Receive:

  • Async FastAPI + Qdrant microservice containerized with Docker Compose.
  • Live endpoints to upload, index, and delete PDF/TXT files.
  • Built-in HTML Demo Chat UI + Swagger docs to visually test uploads and citations (custom frontend design not included).
  • Setup guide & 2 adjustments within scope.


Send me a message with a sample document to get started!

Get to know Mohamed Hany

Mohamed Hany

Ai Engineer

  • FromEgypt
  • Member sinceMay 2021
  • Languages

    English, Arabic
I am a software developer specializing in Artificial Intelligence and backend architecture. I focus on building robust, scalable, and highly optimized solutions using Python, FastAPI, and Streamlit. I have hands-on experience developing end-to-end, context-aware RAG pipelines using ChromaDB and building clean web interfaces with Streamlit. Whether you need an intelligent backend integration, efficient vector database architecture, or a reliable AI chat API from scratch, I deliver clean, maintainable, and well-documented code that solves complex problems.