At Appedology (Pvt) Ltd, I engineer Agentic AI and Graph-RAG pipelines, including a medical coding platform that turns clinical documents into structured ICD-10, CPT, HCPCS, and modifier codes. I’ve reduced LLM prompt context by 30–40% and inference costs by 25–35% through retrieval and context management.
I also built an AI Medical Billing Agent: a LangGraph controller coordinates six specialized GPT-5 sub-agents to take clinical documents through coding and claim drafting. The system reached a 98.5% processing success rate on valid inputs.
Earlier, I developed machine-learning solutions for real-estate valuation at Luckyhiya Associates Pvt. Ltd., improving model accuracy by 15–20%, and created Power BI and Excel dashboards. I work across the AI lifecycle with Python, TensorFlow, PyTorch, FastAPI, and Django, and I’m pursuing an MS in Artificial Intelligence at FAST-NUCES.
