Developed a dynamic Celery queuing enhancement for the etalab-ia/OpenGateLLM repository, enabling running Celery workers to accept new model queues without downtime. This approach improved throughput and resource utilization for model serving by allowing more flexible task routing. The work involved backend development using Python and integrated Redis for queue management. Additionally, the developer simplified Docker configuration, reducing deployment complexity and accelerating environment setup. Unit test coverage and reporting were expanded to strengthen maintainability and performance validation. Collaboration with other contributors was evident through co-authored commits, reflecting a shared focus on code quality and robust, production-ready engineering practices.