A presentation at AI_Dev in in Amsterdam, Netherlands by Oleg Šelajev
Building AI workflows: from local experiments to serving users Oleg Šelajev, Docker
Agentic applications need three things Models Tools Code
Large Language Model github Github Tool User input or event notion Agent Application Notion tool Response SQL tool Internal DB
The Docker Model Runner Run models next to your other containerized services using the tools you’re already using compose.yaml models: gemma3: model: ai/gemma3:4B-F16 services: app: models: gemma3: endpoint_var: OPENAI_BASE_URL model_var: OPENAI_MODEL
The MCP Catalog Run MCP servers using containers without worrying about runtimes or installs anymore
The MCP Gateway Run containerized MCP servers safely and securely directly in your application stack compose.yaml services: mcp-gateway: image: docker/mcp-gateway:latest use_api_socket: true command: - —transport=sse - —servers=duckduckgo - —tools=search,fetch_content app: … environment: MCP_ENDPOINT: http://mcp-gateway:8811/sse
github Large Language Model Github Tool notion User input or event Notion tool Agent Application MCP Gateway Response SQL tool Internal DB
Compose for agents Build and run AI agents using Docker Compose
Cloud Run and Docker Compose Deploy your compose.yaml directly to Cloud Run cloud.google.com/blog/products/serverless/cloud-run-and-docker-collaboration
Everyone can throw together an LLM, some MCP tools, and a chat interface, and get an AI assistant we could only dream of a few years back. Add some “business logic” prompts, and you get an AI workflow; hopefully a helpful one. But how do you take it from a local hack to a production application? Typically, you drown in privacy questions, juggle npx commands for MCPs, and end up debugging OAuth flows before it hopefully starts to make sense.
In this session, we show a repeatable process for turning your local AI workflow experiments into a production-ready deployment using containerized, static configurations.
Whether you prefer chat interfaces or replace them with application UIs, you’ll leave with solid ideas for going from a cool demo to real applications without the existential dread of DevOps.
The following resources were mentioned during the presentation or are useful additional information.
for free. You
can too.