Summary
Ke-chi Chang is a Senior Software & AI Engineer with over six years at MediaTek and a decade of hands-on experience building AI-driven imaging and vision systems for smartphones, tablets, and surveillance products. He specializes in end-to-end AI algorithm development—covering denoising, HDR, demosaicing/remosaic, multi-object/landmark detection, and segmentation—and in tuning those models to run efficiently on ISP and chipset accelerators. His background bridges academic research (co-author of an ECCV paper on camera-aware noise models) and product engineering, enabling rapid translation of novel methods like NeRF and diffusion models into deployable features for clients such as OPPO, Vivo, Xiaomi, and Honor. Ke-chi regularly leads and participates in technical reading groups, keeping production projects aligned with cutting-edge research while collaborating with hardware and QA teams to ensure robustness across diverse scenarios. He brings practical systems experience across sensor fusion, SLAM, and 3D pose work from internships, which informs a pragmatic approach to real-world noise, latency, and resource constraints. Based in Hsinchu, Taiwan, he is known for turning advanced imaging research into stable, high-quality features optimized for mobile and embedded platforms.
10 years of coding experience
碩士, 資工, 碩士, 資工 at 國立清華大學