Our project aims to simplify the lecture recording experience by creating a system to track the presenter around the room, thereby eliminating the need for a dedicated human recorder. We use dynamic zoom and focus to maintain the highest possible clarity for recording notes, while keeping file size and camera requirements down to minimize costs vs comparable alternatives. Our system employs real-time deep learning inference powered by Xilinx’s FGPA platform, enabling breakthrough efficiency that lets the entire system be contained in a single, easily portable box that frees the camera from the demands of wall power, and allows access even in the most challenging and limited of lecture halls.