"Neuru Volex team on AMD Developer Hackathon: ACT III Hackathon"

Team Idea

We are building an AI-powered predictive maintenance copilot for Track 1 (Intelligent Industry). It analyzes machine sensor data, detects early signs of failure, and uses an open LLM running on AMD GPUs (ROCm + vLLM) to explain the fault and recommend maintenance actions. We’re open to collaborators with skills in frontend development, ML, or data engineering.

Team Members

Shahzaman Jatoi

shahzamanghous24709

Submission

TwinMind AI

TwinMind AI is a digital-twin maintenance copilot. It classifies motor faults from vibration data, explains each alert in plain words, and uses an AMD GPU with PyTorch and ROCm to screen sensor data up to 13.8x faster than a tuned CPU.