Computer vision applications are essential for demonstrating practical applications and capabilities of advanced algorithms for real-world use. These demonstrations must run smoothly to effectively demonstrate tasks like position estimation, image segmentation, and object detection. If a demonstration runs smoothly and efficiently, it makes the technology look reliable and will attract people to use and implement it. We have used a combination of advanced techniques and tools to optimize computer vision applications for performance. Profilers are used to identify and address any bottlenecks in the code, ensuring maximum performance. Leveraging hardware accelerators like GPUs, we are able to handle more complex calculations faster. Additionally, we optimize our code by using efficient libraries and frameworks like OpenCV and PyTorch. By optimizing the input data and pre-trained model, we have improved the performance of our application without losing key features. In this project, we have implemented these optimization strategies and tools to great effect. By integrating profiling tools, hardware acceleration, and code optimization, we have significantly improved the application's reliability and performance. Our results show reduced latency and increased framerate on the scale of over 10 times its previous values, providing a clear and smooth user experience. These improvements to our demonstration show that our computer vision technology is effective and ready to provide users with an accurate representation of what computer vision offers.