• Developed predictive algorithms to understand customer behavior with supervised and unsupervised learning techniques using Python libraries including Tensor Flow, Scikit-learn, and Numpy• Implemented YOLO models (YOLOv5, YOLOv7 & YOLOv8) & Darknet-Caffe models using PyTorch and Tensorflow to train 100+ custom image datasets to detect unsafe workplace activity for a client in construction sector. Helped improve model training accuracy by 20% and testing accuracy from 40% to 60% - for mAP@0.5 - 50% and mAP@0.5-0.95 - 50-75% using LSTM & Hyperparameter Tuning• Created models providing insights on customer insights patterns and trends, boosting customer engagement by 3%• Develop Computer Vision models for video analytics safety solutions designed to automatically identify anomalous actions in real time in high risk environments• Data Engineering which consisted of Designing and building data pipelines, Developing and maintaining data warehouses and data lakes, Integrating data from different sources, Cleaning and transforming data, Ensuring data security and compliance, Troubleshooting data problems• Performed Data Analytics using Pandas & Matplotlib to create Statistics Tables on live data gathered for a day, from different cameras in construction sites in Singapore, Saudi Arabia for object detection and activity recognition work for a client project