During December 2024, Mikhail Lhakov developed deployment infrastructure for the embedding service in the aimclub/ProtoLLM repository, focusing on reproducibility and scalability for SDK users. He designed Docker Compose configurations and deployment scripts to enable one-command setup of the embedding service with ChromaDB and HuggingFace models. By integrating supporting services such as RabbitMQ, Redis, Celery workers, and Flower, he established a robust environment for scalable task processing and monitoring. His work leveraged Dockerfile and YAML, applying CI/CD and containerization best practices. This foundational engineering effort addressed onboarding challenges and set the stage for reliable future feature development within ProtoLLM.