class KamranUmer:
role = "AI-Backend Engineer @ CCRIPT Agency"
experience = "4+ years"
location = "Islamabad, Pakistan 🇵🇰"
builds = ["LLM apps & RAG", "AI agents", "computer vision", "FastAPI / Flask backends",
"cloud deployments"]
learning = "robotics & mechatronics → github.com/KamranUmer/ROBO-TOBO"
def say_hi(self):
return "I turn AI models into products people can use. Now I'm teaching them to move. 🤖"|
Learning mechatronics in public: every chapter of Mechatronics System Design turned into Python simulations, then built on real hardware. |
RAG chatbot that recommends doctors in Pakistan by city, condition, budget and gender preference. |
|
Audience analysis, Gemini-generated posts and images, and performance tracking in one dashboard. |
Custom-trained detector for medical X-rays, mAP@0.5 = 0.914, with a YOLO-NAS version and a Streamlit app. |
|
Plant disease detection from leaf photos (96.7% validation accuracy) with an AI treatment assistant. |
Retrieval-augmented QA over 50,000 Wikipedia passages: MPNet retriever, Pinecone, BART answer generator. |



