The Event Loop & Appcharge

This talk covers how we diagnosed a slow production flow by breaking the request into its real stages, from platform and microservice down to serialization and response write time and using an AI agent to investigate each one: navigating code, forming hypotheses, generating CloudWatch queries, and reading the graphs. The takeaway for backend developers: AI isn't just a code writer. It's a sharper way to ask the right questions. Agentic SDLC Graph | by Ron Mizrahi R&D Team Lead @ Getruck AI can already write the code, but a human still drives the process, step by step, prompt by prompt. The next move is to build the development process itself as a graph: nodes, edges, guards, shared state, loops, and predefined checkpoints where a human can step in. The LLM keeps the flexibility and judgment inside each step, while the graph gives the SDLC structure, control, and observability. Live demo: from an idea all the way to an open PR. Please note that registration does not guarantee admission. Attendance is subject to host approval. You will receive a separate confirmation email once your registration has been approved.