Inspiration
We wanted to create a system that demonstrates how multiple AI agents can collaborate dynamically, just like a distributed team in the cloud. Our goal was to build a framework that’s both scalable and easy to extend — something developers can adapt for orchestration, automation, or AI coordination.
What it does
The project features a master agent that decomposes and routes tasks to multiple worker agents, each with its own role. It supports dynamic registration, so new workers can join automatically without code changes.
How we built it
We used FastAPI for APIs, Redis as shared memory, and Docker Compose for deployment. Each agent runs independently, communicating through REST endpoints and Redis.
Challenges we ran into
Ensuring smooth communication between multiple containers and handling asynchronous task routing was tricky. We also needed to design a clean interface for worker registration and metrics.
Accomplishments that we're proud of
We built a functional, modular, and easily scalable multi-agent system that can grow without touching the master’s logic.
What we learned
We learned how to design distributed systems with modular components, apply clean API design, and manage inter-agent synchronization using Redis.
What's next for gastroguide
We plan to add worker auto-scaling, message-queue routing, and LLM-powered decision modules for intelligent orchestration.
Log in or sign up for Devpost to join the conversation.