Adamou Zakari@zadamou3224221

This project is a Hybrid Token-Efficient Routing Agent built for Track 1 of the AMD Developer Hackathon: ACT II. The core idea is simple: not every task needs the most expensive AI model to be answered correctly. Sending every request to a powerful model wastes tokens and money, so this agent analyzes each incoming task and decides, in real time, which Fireworks AI model is the cheapest one capable of answering it accurately. The routing logic detects whether a task is code-related (debugging, function writing, syntax errors) or general-purpose (facts, reasoning, summarization, sentiment, NER). Code-related tasks are routed to Kimi K2.7 Code, a model specialized for programming tasks. All other tasks are routed to MiniMax M3, a lower-cost general-purpose model. Both models are used exclusively through the Fireworks AI API, as required by the hackathon rules. The agent was tested against all eight official capability categories (factual knowledge, math, sentiment classification, summarization, named entity recognition, code debugging, logical reasoning, code generation) and produces clean, direct answers in English. The entire application is containerized with Docker, reading tasks from /input/tasks.json and writing structured answers to /output/results.json, matching the hackathon's standardized evaluation environment. The image is published publicly on Docker Hub at adamou1/amd-routing-agent, well under the 10GB size limit at 47MB.

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.