- Trained and implemented YOLOv8 models on sport-related analytics to detect players' equipment, segment for team assignation, key points for posture, and post-process for action analysis- Set up and configured Nvidia Orin systems using Nvidia SDK Manager in Linux and set up a virtual environment to avoid errors in project dependencies- Establishing a reliable TCP/IP connection to ensure smooth communication between the production computer and Nvidia Jetson Orin for model inference.- Analyzed model inference times, and identify which models give an improvement in inference time while retaining inference accuracy.

- Assisted in the frontend development of a React-based web application prototype intended to ease the process of searching for company-wide project-related documents as part of a data centralisation initiative, aiming to improve accessibility and accuracy of internal data. - Developed automated forms with Microsoft PowerApps, helping to reduce calculation errors of CAPEX and OPEX compared done manually. The form also automated the process of sending approval emails to the relevant superiors, ensuring efficient communication and streamlined approval workflows.