"ModelMargin AI team on AMD Developer Hackathon: ACT II Hackathon"

Team Idea

ModelMargin AI cuts AI inference costs using AMD GPU-powered model routing.ModelMargin AI cuts AI inference costs using AMD GPU-powered model routing.

Team Members

Shone Muune

ShoneKahuria

Software Developer

Submission

ModelMargin AI — Profit-Aware Inference

Turn AI prompt logs into a profit plan: route workloads to cheaper AMD-hosted Gemma inference with quality guardrails, and prove the savings in a Margin Dossier — measured live on AMD Instinct MI300X.