AI & Automation · Enterprise
Multi-Agent AI Workflow System
A role-based agent team — researcher, writer and editor — coordinated through LangGraph with layered memory that retains context across sessions and learns from historical task results.
CrewAILangGraphChromaDBFastAPICloud Run
3-role
Agent team topology
3-tier
Short, long and persistent memory
What we built
01
Agent architecture
Agent roles and state schema built on LangGraph with conditional routing.
02
Task delegation
CrewAI goal-driven definitions distributing work across the agent team.
03
Vector memory
ChromaDB integration providing cross-session context retention.
04
Human-in-the-loop
Cloud Run deployment with full CI/CD and human oversight at decision points.
Next case study
AI Voice Assistant Deployment