I’m building a multi-agent AI system designed to streamline greenfield software projects. This is **not** a code-generation tool. The focus is on: - Requirement analysis - Task breakdown - Assignment & tracking - Progress reporting - Surfacing blockers and risks ## Core Requirements - Agents deep dive into a project idea by asking questions - Once the agent understands exactly what the user wants, appropriate documents are created and handed over to the next agent ## Tech Currently using **Google ADK** and having a semi working flow. But open to **LangChain**, **Autogen**, or alternative orchestration frameworks that better support: - Multi-agent coordination - Memory (short- and long-term) - Tool access & integrations --- ## Deliverables - Working demo project with clearly defined agents - Architecture diagram - Setup documentation - Extendable agent design --- Looking for someone experienced in multi-agent systems, orchestration architecture, memory design.