Maneesha Herath@maneesac432163060

Our Track 2 agent captions short videos in four tones: formal, sarcastic, humorous-tech, and humorous-non-tech. It reads /input/tasks.json, downloads each clip, samples key frames with ffmpeg, transcribes audio with Whisper when present, and builds a factual JSON context with a Fireworks vision model. A text model writes a neutral base caption, then rewrites it into the four styles with few-shot prompts. A judge step checks accuracy and tone and regenerates once if needed. The container writes /output/results.json and exits. Perception can run on CPU/GPU; generation uses Fireworks. We keep facts shared across styles so humor does not invent details. Built for varied scenes beyond the three sample clips.

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.