Hira Sardar

PROFESSIONAL PROFILE Applied AI and Machine Learning practitioner with hands-on experience in computer vision, video analytics, and deep learning systems. Skilled in building and benchmarking scalable AI pipelines on real-world datasets, with a strong focus on translating research into production-ready solutions. EDUCATION Bachelor of Science in Computer Science (BSCS) NUST School of Electrical Engineering and Computer Science , Islamabad , 3.77/4.0 (2026) INTERNSHIP EXPERIENCE VisionRD 01-Dec-2025 - 24-Jan-2026 Worked on large-scale egocentric video datasets for Factory Copilot use cases, applying and benchmarking state-of-the-art AI models including BAFormer, SAM2, and V-JEPA for video understanding and representation learning. Designed and evaluated task- specific pipelines for industrial scenarios, enabling reliable video-based analysis across multiple factory workflows. Zhejiang University 30-Jul-2025 - 31-Aug-2025 Conducted experimental analysis of state-of-the-art video watermarking techniques on biometric video datasets, evaluating their impact on rPPG-based heart rate estimation. Designed and benchmarked robust pipelines to assess signal degradation and improve model reliability under real-world video transformations. TUKL–NUST Deep Learning Lab & National Centre of Artificial Intelligence (NCAI) 01-Jun-2023 - 31-Jan-2025 Worked on two applied deep learning projects: Interactive Brain Education via Augmented Reality and Plant Disease Diagnosis via Co-Salient Object Detection. Designed high-fidelity 3D brain models and interactive visualizations using Blender, integrating deep learning and computer vision concepts to support AR-based neuroeducation and evaluating feasibility for real-world deployment. Implemented and benchmarked multiple co-salient object detection models on custom agricultural datasets to localize disease- affected regions in plant imagery. Optical Networks and Technologies (ONT) Lab 01-Jun-2024 - 30-Sep-2024 Built and optimized machine learning pipelines for signal quality prediction using large-scale tabular network performance data. Applied and compared advanced techniques including active learning, knowledge distillation, federated learning, and explainable AI to improve model performance and reliability in real-world network scenarios. Pakistan Science Foundation – Science Talent Farming Scheme (STFS) 01-Jan-2021 - 31-Dec-2022 Selected for a highly competitive national research program under the Science Talent Farming Scheme (STFS), where I conducted a mentored project titled “Investigation of Cubic B-Spline Scheme for Modeling of Curves,” applying mathematical modeling and data analysis techniques. Gained hands-on experience in structured problem-solving, experimental design, and analytical thinking through inquiry-based research training. FINAL YEAR PROJECT

Applied AI and Machine Learning practitioner with hands-on experience in computer vision, video analytics, and deep learning systems. Skilled in building and benchmarking scalable AI pipelines on real-world datasets, with a strong focus on translating research into production-ready solutions.