Yash Sud

Data Scientist and AI Engineer with an M.Sc. in Artificial Intelligence and hands-on experience building, training, and deploying deep learning systems for computer vision and image restoration. Skilled in optimizing YOLOv8 and CNN architectures, engineering large-scale preprocessing pipelines, and delivering production-ready, measurable results. Seeking to apply applied ML and computer vision expertise to solve high-impact, real-world problems.

AI/ML Intern - CSIR – CRRI (Central Road Research Institute)

(2025-01 - 2025-06)

• Designed, trained, and optimized a custom YOLOv8-based computer vision model for road marking and traffic sign detection, achieving 92.4% mAP@50 and 45 FPS real-time inference under the guidance of Dr. Ch. Ravi Sekhar (Chief Scientist, Transportation Planning & Environment).
• Built an interactive Tkinter GUI to visualize bounding boxes, display confidence scores, and export annotated outputs, cutting manual validation time for the research team by 40%.
• Preprocessed and annotated 2,000+ traffic images and videos, applying augmentation techniques (lighting, blur, noise) that measurably improved model robustness across environmental conditions.

Internship - AD Infocom Systems - Remote

(2025-01 - 2025-07)

• Developed and optimized a CNN-based Image Colorization model in LAB color space using TensorFlow and Keras, automating grayscale-to-color image restoration for media archiving and achieving a PSNR of 24.5 dB.
• Engineered custom preprocessing pipelines for 5,000+ high-resolution training images, reducing training latency by 25%.
• Delivered a production-ready batch-processing restoration model, cutting manual touch-up time for historical media by 30%.

M.Sc. - Artificial Intelligence - Central University of South Bihar (2023 - 2025)

Bachelor of Computer Application - Patliputra University (2020 - 2023)