Electromagnetic interference (EMI) remains a major bottleneck in power converter design, often requiring costly expert reviews and lengthy measurement campaigns. We propose an AI-powered EMI diagnosis agent that combines a fine-tuned vision–language model (VLM) with retrieval-augmented generation (RAG), memory, and tool integration. Trained on PCB layouts, netlists, and EMI measurements, the agent can rapidly analyze designs, retrieve relevant knowledge, and deliver clear, explainable recommendations. This approach reduces reliance on manual review, accelerates design iterations, and provides a scalable solution for EMI compliance in industrial workflows.
.jpg "Yirui Yang (he/him/his) photo")
.jpeg.jpg "Shuo Wang photo")