Sai Gowtham Jayavaram@saigowtham69978409

This project implements a Hybrid Token-Efficient Routing Agent that autonomously decides which Fireworks AI model is the cheapest one capable of answering each task accurately. Rather than routing every task to a single powerful model, the agent classifies each incoming prompt by difficulty — easy, medium, or hard — using keyword detection and routes it to the appropriate model tier. Easy tasks are handled by a lightweight cheap model, while medium and hard tasks escalate to a stronger model. If the cheap model produces an invalid or insufficient output, the agent automatically escalates and retries. This approach minimizes total Fireworks token consumption while maintaining 100% accuracy across all tasks, directly optimizing for the hackathon scoring criteria of lowest tokens with highest accuracy.

To participate in the hackathon, simply click Sign up with AMD. If you are not already a member of the AMD AI Developer Program (ADP), you will need to create an account. ADP members can access AI tutorials, experts, and community support. In addition, new members can claim a free month of DeepLearning.ai pro, $100 in AMD GPU cloud credits, and $50 in Fireworks AI API credits. All hackathon participants will receive on hackathon launch additional compute and API credits for their projects. Ready to build what's next on AMD? Sign up with AMD below to secure your spot.