Dhina

Machine Learning Engineer with 6 months of industry experience specializing in Computer Vision, Deep Learning, and Edge AI. Built production-ready AI systems including YOLOv8 PPE Safety Monitoring (90%+ mAP, 25+ FPS), AI-powered visual inspection, and OCR automation solutions. Proficient in PyTorch, OpenCV, TensorRT, CUDA, FP16/INT8 quantization, and deployment on NVIDIA Jetson for real-time industrial AI applications.

ML DATA Intern – Computer Vision - SeeWise AI – LMBees Pvt. Ltd. - Chennai, India

(2026-03)

• PPE Safety: Developed a YOLOv8-based multi-camera CCTV PPE detection system achieving 90%+ mAP across 6 classes at 25+ FPS; reduced manual monitoring effort by 60% through real-time alerting.
• Model Training: Trained and optimized YOLOv8 models on 3,000+ industrial images over 200+ epochs, leveraging data augmentation, hyperparameter optimization, and failure-case analysis to improve model robustness.
• Data Pipeline: Developed scalable dataset preparation pipelines using CVAT and Locate Anything 3B, improving data preprocessing efficiency by 40% and accelerating model development cycles.
• OCR Pipeline: Built OCR-based inspection pipelines capable of extracting and validating product labels and serial numbers with 95%+ text extraction accuracy for automated quality verification.
• Edge AI Deployment: Optimized and deployed YOLOv8 models on NVIDIA Jetson edge devices using TensorRT, enabling low-latency real-time inference for industrial computer vision applications.
• Model Optimization: Applied FP16 and INT8 quantization, TensorRT optimization, and CUDA acceleration to reduce inference latency, improve throughput, and optimize GPU memory utilization for edge deployment.

B.Tech - Computer Science Engineering (Data Science) - Hindustan Institute of Technology and Science (2026-08)