Junsik Youn
I am interested in vision processing and efficient execution of vision language AI in resource-constrained systems. My research focuses on Vision Language Model (VLM), On-Device AI, and hardware/software co-design for accelerator-based AI computing. I am particularly interested in building a system architecture that can reliably and run visual and multimodal workloads on edge devices with low latency. My goal is to connect computer architectures with real-world AI applications through efficient and practical system-level optimization.