This session explores practical strategies for deploying computer vision AI on far-edge devices under strict resource constraints. While highlighting FPGA-specific strengths, such as customizable dataflows, fine-grained quantization control and efficient near-sensor processing, we will also cover techniques applicable across edge platforms. We will address challenges like limited model capacity and memory, and discuss pipeline design, quantization methods and hardware-aware training workflows that preserve gradient flow and feature quality—enabling robust, real-time AI performance even in highly constrained environments.
Efficient Computer Vision at the Far Edge: Design and Training Under Constraints
Session Speakers
-
Nicolas Widynski
Co-Founder and CEO, Roboflow
Nicolas Widynski is a Fellow at Lattice Semiconductor, where he drives the company’s AI and computer vision initiatives. Nicolas joined Lattice in 2021 following its acquisition of Mirametrix, where he was part of the founding team and led development of advanced vision-based technologies. He brings deep expertise in computer vision and machine learning, supported by academic research and industry experience. Before joining Lattice, Nicolas held two postdoctoral positions at the University of Montreal in computer vision and medical imaging. He earned his PhD in Computer Vision from Télécom ParisTech (now Télécom Paris) and his MSc from Télécom ParisTech/University Pierre et Marie Curie.
