At Lionic AI Labs, I design and deliver computer vision, NLP, and deep learning solutions for client automation and predictive analytics. I take projects from requirements and data preparation through model evaluation and software integration.
I developed LangGraph-based agentic AI workflows that coordinate agents through task planning, dynamic tool calling, and persistent memory. I also led development of a live-camera employee monitoring system that uses face detection, tracking, and video processing to generate presence and activity signals.
I built a multi-document RAG pipeline with Hugging Face, LangChain, and Qdrant, using dense and sparse retrieval with cross-encoder reranking. My projects also include a local-LLM PDF chatbot designed to avoid API costs and keep document data private.
My other work spans real-time edge video analytics with YOLOv8, ByteTrack, and TensorRT, as well as PyTorch projects in facial emotion recognition and crop disease classification. I’m pursuing an MS in Artificial Intelligence at Lahore University of Management Sciences (LUMS).
