C# or Python AI Frame Detection Engine

We are looking for an experienced Python/C# developer with strong Computer Vision and Machine Learning experience to build a real-time sports video detection engine. This is a serious development project that will eventually be integrated into a much larger application. We are looking for someone who understands computer vision, object detection, OCR, neural networks, APIs, and real-time video processing. What the System Needs to Do Our system will continuously send screen video frames to your backend engine through an API. The engine needs to identify and return: Sport Detection Football Basketball Rugby American Football Other sports as training data becomes available Logo Detection Identify detected teams, leagues, and competition logos Examples: Real Madrid, Bayern Munich, Liverpool, Inter, UEFA, etc. Player Identification Identify players based on their shirt number and available training data Example: 10 - Kylian Mbappe, 11 - Rodrygo Player names and numbers will be provided through text files. Home / Away Kit Detection Determine whether the detected team is wearing its home or away kit. Example: Real Madrid Home / Real Madrid Away. Match Information Read team names and scores displayed on the screen. Extract information from the scoreboard when available. Ball Detection / Possession Detect the ball. Determine which identified player is currently holding or controlling it. Training Data We will provide the sports data and training images in a structured folder system similar to: SportsDetection\Football\2026\Ball\ SportsDetection\Football\2026\Logos\Arsenal\ SportsDetection\Football\2026\Wear\Arsenal\Home or Away\ The system should be designed so additional sports, teams, players, kits and training data can be added without having to rebuild the entire application. Backend / Worker System You may use: Python + FastAPI, or C# + ASP.NET Core The system will use multiple workers to process incoming frames. We will initially run 3 workers, but the architecture needs to allow the number of workers to be increased as the system scales. Workers should be reusable/re-initializable, and the engine should maintain the identified information in memory and return the results to the frontend while frames are continuously being processed. Hardware Requirements The solution needs to work in: Windows Linux GPU environments Non-GPU/CPU environments Performance should be considered carefully because the system will be processing continuous video frames. Frontend / Demo You will also need to create a simple frontend/dashboard demonstrating that the system works. The dashboard should show the detected information in real time, such as: Sport Teams/logos Players Shirt numbers Home/Away Match information Score Ball detection Current player possession The frontend does not need to be visually complex. The main purpose is to demonstrate the backend engine working correctly. Security & Reliability This system will eventually be integrated into a much larger application, so the code needs to be: Secure Reliable Well structured Maintainable Properly validated Resistant to unexpected input Suitable for continuous operation We will perform security and reliability testing against the application before accepting the final delivery. Deliverables The successful developer will provide: Complete source code Backend/API Computer vision/ML engine Worker management system Simple frontend/dashboard Training/data integration GPU and CPU support Windows-compatible implementation Setup/documentation instructions Demonstration of the complete system All source code and intellectual property The code will be integrated into a larger application, so clean and well-organized source code is essential. Important Please bid your actual price. Do not submit an artificially low bid simply to start a conversation. Your bid will be treated as your actual project budget. Payment will be made once the agreed system has been completed, demonstrated, and confirmed to work on our machine. Requirements You should have proven experience with: Python C# / ASP.NET Core Computer Vision Machine Learning Neural Networks Object Detection OCR Image/Video Processing REST APIs Real-time video processing GPU/CPU optimization Experience with sports video analysis, player recognition, team/logo detection, ball tracking, or similar computer vision systems is highly preferred. When Applying Please provide: Your relevant experience with Computer Vision / ML. Examples of similar projects you have completed. Your portfolio or GitHub if available. Which technology you would recommend: Python/FastAPI or C#/ASP.NET Core, and why. Your proposed architecture for the worker system. Your estimated delivery time. Your fixed project price. We may conduct a short interview before awarding the project. A working demo will also be required through Upwork before final acceptance. There may be additional development work for the successful developer after the initial project. We are looking for someone who can take ownership of the technical side of the project and build a solid foundation that can be expanded as more sports, teams, players and data are added.