Browse categories
Explore
Fiverr Pro
English
$
USD
I will build a rag chatbot with custom knowledge base using python
Muhammad H
About this gig
I'll build a production-ready RAG (Retrieval-Augmented Generation) chatbot that answers questions from YOUR data documents, websites, or databases using Python, LangChain, and OpenAI or Claude.
What I deliver:
Document ingestion pipeline PDF, DOCX, TXT, HTML, CSV, or web pages
Smart chunking and embeddings with Pinecone or pgvector
Accurate answers with source citations to reduce hallucinations
Conversation memory for natural multi-turn chat
FastAPI backend with clean REST endpoints
Admin endpoint to upload and update your knowledge base
Docker-ready code with AWS deployment support
I've built RAG systems used in real products, including a multi-tenant enterprise RAG chatbot (MeiChat) serving multiple clients from custom knowledge bases.
Technologies: Python, LangChain, LangGraph, OpenAI GPT-4o, Claude, Pinecone, pgvector, PostgreSQL, Redis, FastAPI, Docker, AWS
Whether you need a simple document Q&A bot or a full production system with citations, memory, and monitoring I deliver clean, scalable code that works.
Get to know Muhammad H
Muhammad H
AI Engineer specializing in LLM Agents RAG Systems Python and FastAPI
- FromPakistan
- Member sinceJul 2026
Languages
English, Urdu, Punjabi
I build production AI systems - AI agents, RAG pipelines, LLM applications, and multi-tenant SaaS backends - using Python, FastAPI, LangGraph, and PydanticAI. Over 5 years I have shipped agentic assistants with tool calling and real-time streaming, multi-tenant RAG chatbot platforms with Pinecone and OpenAI/Claude integrations, and scalable backend systems on AWS with PostgreSQL, Redis, and Docker. I work with startups and product teams who need AI that works in production, not just demos. Let's connect.
FAQ
What data sources can the chatbot use?
PDF, DOCX, TXT, HTML, CSV files, websites, and databases like PostgreSQL. I build an ingestion pipeline that chunks and embeds your content into a vector database such as Pinecone or pgvector.
Will the chatbot make things up (hallucinate)?
I ground every answer in your data using retrieval and add source citations. When the answer is not in your knowledge base, the bot says so instead of guessing. I also tune prompts and retrieval settings to maximize accuracy.
Can you deploy it and integrate it with my website?
Yes. I deliver Docker-ready code, can deploy to AWS, and expose a clean REST API you can connect to any website or app. I will send a short questionnaire after you order to gather your data sources and requirements.
