Integrating three neural networks into a single automated workflow presented multiple challenges. There was a need to perform additional calculations and data transformations between each network, while also ensuring robustness and scalability. Moreover, one of the tools was restricted to Windows, and the entire calculation process was time-consuming. To address these issues, we separated the system into two distinct components: a dedicated worker for Windows and a lightweight backend compatible with any OS. Leveraging a task queue, we ensured reliable task delivery and effective load balancing. By adopting an event-driven approach, we streamlined the flow, achieving seamless integration between any number of workers and the backend, enhancing the system's overall extensibility. Building the web app took about 3 months. The client was completely satisfied with the result, especially highlighting the speed of execution, constant communication and feedback, and flawless work ethic. Despite the preconceptions of other agencies that it’s almost impossible to build such a product, our client believed in us and entrusted us with the implementation of this project, and we are very grateful for their trust and proud of ourselves for its successful realization.