"Omega team on AMD AI League Hackathon"
Team Idea
A completely private, ambient meeting intelligence tool that runs continuously in the background during calls. It handles live multi-speaker transcription, real-time factual lookups, and instant action-item generation without sending proprietary audio over the internet. Tech Stack: Model: Whisper (quantized via INT8/FP16) for speech-to-text paired with a fast reasoning LLM (e.g., Llama 3.1 8B or Mistral 7B). AMD Advantage: Parallelized audio pipeline utilizing AMD NPU/GPU execution threads.
Team Members
Omuga1
omuga1612
Submission
No submission yet