01
The Problem
Programmatic ad buying is reactive. Teams manually set bids and budgets with no real-time intelligence to optimize across multiple DSPs simultaneously, leaving significant performance on the table.
02
What It Does
A multi-agent system on LangGraph that runs parallel agents for bid recommendation, auction simulation, and performance optimization across Google, Meta, Amazon, TTD, and LinkedIn, all in real time.
03
Why I Built This
AdTech operates at massive scale but still relies on rules-based automation. I wanted to explore what happens when you replace that with a reasoning system that can weigh tradeoffs like a strategist at machine speed.