Muhammad Ahad Hassan Khan

PROFESSIONAL PROFILE I am a Computer Science undergraduate at NUST with strong research and applied experience in machine learning, computer vision, and LLM-based systems. My work spans end-to-end AI pipelines, including model design, dataset preparation, rigorous evaluation, and deployment, with applications in precision agriculture, real-time analytics, and policy-aware information systems. I have contributed to research projects involving multispectral vision, retrieval-augmented generation, and generative models, and I am particularly interested in developing reliable, interpretable, and scalable AI systems for real-world use. I aim to continue learning, enhance my technical skills, collaborate on impactful projects, and build robust, industry-ready AI systems, while also pursuing graduate studies to deepen my knowledge and expertise in artificial intelligence and computer science. EDUCATION Computer Science SEECS , Islamabad , 3.57 (2026) INTERNSHIP EXPERIENCE Siber Koza - CENTAIC (Centre of Artificial Intelligence and Computing) 23-Jun-2025 - 23-Aug-2025 Computer Vision & NLP Intern, Siber Koza — CENTAIC, NUST Gained hands-on experience across computer vision, generative AI, and NLP systems through applied research and prototyping. Worked with synthetic data generation using BlenderProc, explored diffusion models including Stable Diffusion with LoRA/DreamBooth fine-tuning, and designed an ALPR system for Pakistani license plates using YOLOv11 optimized with NVIDIA DeepStream for real-time analytics. Built retrieval-augmented generation (RAG) applications using LangChain and Streamlit for structured and unstructured data, and implemented stereo depth estimation using epipolar geometry on the KITTI dataset. Developed multi-agent LLM workflows using LangGraph and gained practical exposure to cloud-based AI pipelines on AWS. School of Electrical Engineering & Computer Science - SEECS 20-Mar-2025 - 19-Jun-2025 Research Assistant (Computer Vision), Agri-Drone Project — SEECS, NUST Worked on an end-to-end agri-drone vision pipeline for crop localization and growth stage prediction using RGB and multispectral imagery. Modified the Ultralytics YOLOv11 framework to support 4-channel RGB+NDVI input and multi-task learning, enabling simultaneous crop detection and maturity estimation. Developed a UNet-based CNN to approximate NDVI from RGB drone images and validated it against multispectral ground truth. Led dataset annotation and preprocessing for 1,500+ drone images, including multispectral band alignment and expert-verified growth stage labeling. Deployed a real-time inference pipeline via a FastAPI service integrating NDVI estimation, detection, and visualization, optimized for efficient deployment. Optical Networks & Technologies Lab - NUST 10-Jun-2024 - 03-Sep-2024 Deep Learning Intern, Optical Networks and Technologies Lab — SEECS, NUST Conducted applied research in supervised and deep learning for optical network performance modeling. Developed GSNR prediction models using regression, tree-based methods, and neural networks, optimized through regularization, grid search, early stopping, and dropout. Applied transfer learning across geographically distinct datasets and leveraged active learning to reduce labeling costs. Implemented knowledge distillation for efficient model compression and explored federated learning for privacy-preserving training across distributed data sources. Employed explainable AI techniques (SHAP, LIME, PDP) to interpret model behavior and analyze feature importance, strengthening experimental rigor and evaluation practices.

I am a Computer Science undergraduate at NUST with strong research and applied experience in machine learning, computer vision, and LLM-based systems. My work spans end-to-end AI pipelines, including model design, dataset preparation, rigorous evaluation, and deployment, with applications in precision agriculture, real-time analytics, and policy-aware information systems. I have contributed to research projects involving multispectral vision, retrieval-augmented generation, and generative models, and I am particularly interested in developing reliable, interpretable, and scalable AI systems for real-world use. I aim to continue learning, enhance my technical skills, collaborate on impactful projects, and build robust, industry-ready AI systems, while also pursuing graduate studies to deepen my knowledge and expertise in artificial intelligence and computer science.