Multi-Agent AI System
An orchestration system where three specialized AI agents — Planner, Researcher, and Writer — collaborate to produce content autonomously.
Overview
A demonstration of how specialized AI agents can collaborate to produce better output than any single model call. Built as a portfolio piece targeting AI engineering roles.
The problem
Single-prompt AI workflows hit a wall when tasks get complex.
What I built
Three agents — Planner, Researcher, Writer — each with a focused responsibility, communicating through a structured pipeline. Planner decomposes the task, Researcher gathers context, Writer produces the final output. FastAPI backend, React and Vite frontend, deployed on Render and Vercel.
The outcome
A working demo showing how multi-agent systems handle complex tasks that single prompts can't. The architecture is the real output — clean separation of concerns, clear data flow between agents, and a UI that makes the process visible.