zeeshan.
Start a project
I build the layer between your documents and the answers people ask for.
RAG systems, grounded chatbots and AI agents that answer from your own policies, product data and files, instead of a model guessing.
Five kinds of systems. One rule: answer from real data.
If a plain script or a search box solves your problem better than AI, I'll say so before we start.
RAG & knowledge systems
Documents, policies and internal know-how turned into a system that answers from them.
AI chatbots
Support assistants for websites, WhatsApp and Telegram that stick to your knowledge base.
AI agents
Multi-step workflows that plan, retrieve, verify and write, with a human in the loop.
Document intelligence
OCR, extraction, translation and question answering over scanned or messy files.
Voice AI
Speech to text, retrieval, then a spoken or written reply, running locally if needed.
Systems I've built, start to finish.
Computer science graduate, self-taught into AI systems that ship.
I studied CS at International Islamic University Islamabad and now work as a freelance AI engineer. My focus is narrow on purpose: retrieval-augmented generation and the chatbots and agents built on top of it. Most of what I know came from building real things end to end, from raw documents to a deployed endpoint, in Urdu and English.
The stack I actually build with.
parsing → retrieval → LLM → API → deployA short, honest list. Not a sales pitch.
Have documents that should be answering questions on their own?
Tell me what you're trying to build and what data it needs to know. I'll tell you honestly whether RAG is the right approach before we start.
zeeshanqadir568@gmail.com