@yadikikabeer0229511

TokenHawk is a verification-first, token-efficient AI agent built for Track 1 of the AMD Developer Hackathon. It handles factual questions, mathematics, sentiment, summarization, named-entity recognition, code debugging, logical reasoning, and code generation. For each task, TokenHawk first extracts the required answer format and constraints. It then attempts a local solution using a compact language model, deterministic solvers, and category-specific checks. Answers can be validated through arithmetic recomputation, code execution, format rules, entity checks, and other evidence-based safeguards. Only unresolved tasks are sent to an allowed Fireworks model using compact prompts and controlled outputs. The system runs in a reproducible Linux/AMD64 Docker container and is designed to preserve reliable answer quality while minimizing unnecessary remote tokens and inference cost.

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.