• Extended PwC metaverse platform to include seamless integration of object formats like gltf, glb, and obj 3D models into the system by loading objects directly at runtime so that users can import objects and create their own customization and interact with the platform for an immersive experience.• Designed and implemented RESTful APIs, focusing on CRUD operations, ensuring smooth interaction with Microsoft SQL Server for efficient data management.• Employed optimization techniques resulting in a reduced base project size (2gb to 600mb), facilitating seamless storage and transfer among team members, thereby enhancing collaboration efficiency.• Implemented lazy loading technique for objects, enhancing user experience by loading items only when in the camera view, directly from Azure cloud storage.• Worked on compressing object models to enhance performance and reduce lag, utilizing Unity's optimization features for an improved user experience.• Collaborated on the deployment process by adding endpoints with Azure, facilitating the smooth deployment of the application in WebGL format.• Optimized the application's load time by integrating Content Delivery Network (CDN) solutions and employing various optimization techniques within Unity.

• Created an algorithm for automating the detection of utility poles across diverse counties in the USA, a critical step in fiber deployment planning to assess the viability of aerial deployment, thereby reducing costs.• Leveraged Google Street View APIs to capture images at regular intervals along streets, facilitating the identification of utility poles using advanced object detection models.• Conducted a comprehensive analysis of various object detection models, including Detectron, Detectron2, and YOLO, to identify the most suitable model for utility pole detection.• Implemented hyperparameter tuning techniques to optimize the performance of the selected object detection model, ensuring accurate and efficient utility pole detection.• Engineered an automated image annotation pipeline to streamline the annotation process, improving efficiency by automatically annotating images based on insights gained from previous iterations.