@sumanta_das423

KORA is a token-efficient routing agent for the hybrid local-remote setting of Track 1. Most routing systems optimize the remote call. KORA asks a different question first: should the remote call happen at all? A deterministic decision layer resolves routine work locally, with zero remote inference, and escalates to a remote model only when nothing cheaper is confident. Because the AMD Developer Hackathon scores remote token usage, avoided calls translate directly into a lower score. The routing core is answer-blind: it decides how to handle each unit of work before generating an answer, keeping the decision deterministic and reproducible rather than delegating it to a probabilistic classifier. This makes routing decisions auditable, which we see as the foundation for a broader execution control layer for agentic workflows. For this hackathon we focus on validating the routing tier: deterministic-first deflection, local escalation, and remote inference only when needed.

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.