Muhammad S.

Most "AI expert" profiles here are long on buzzwords and short on proof. Mine isn't. I've delivered 200+ five-star projects across AI Integration, Computer Vision, and AI Agent Development, backed by a published research paper, not just a portfolio page. What I do: AI Integration - LLM and NLP systems wired into real business workflows: RAG pipelines, sentiment analysis, condition-monitoring systems that combine LLMs with data engineering, and AI-powered MVPs that connect models into a client's actual product instead of sitting in a notebook. Computer Vision - my tracking research was published at AIIT 2025, growing out of work at Pakistan's National Center for Physics on long-term object tracking. On Upwork that's shown up as a corn root phenotyping pipeline, a YOLO-based swim performance tracker, and an ongoing YOLOv11 project with a custom loss-function enhancement. I've also built a real-time multi-object tracking system running at 30 FPS. AI Agent Development - I'm currently building production multi-agent systems with LangChain and LangGraph for a financial services client: multi-step, action-taking workflows rather than simple chat responses. If you want serious ML and integration depth brought into your agent build, that's exactly where I am right now. Tech stack: Python, LangChain, LangGraph, OpenAI and Claude APIs, RAG and vector databases, YOLOv8, CLIP, FastAPI, Next.js, Supabase, TensorFlow, and n8n for workflow automation. I'm based in Pakistan, reply within a couple of hours most days, and build a working prototype before I ever send a proposal, so you can see the thing running before you commit to anything. If you need AI Integration, Computer Vision, or an AI Agent built by someone who can show you the receipts instead of just the sales pitch, send me a message.