Perception Research Engineer · Konboi
PhD · Computer Vision & Machine Learning
Deep learning for 3D perception — geometric models, multi-sensor fusion, and uncertainty-aware learning.
Recent work: online LiDAR–camera calibration and uncertainty estimation for autonomous driving.
Konboi · Motional · Toyota · UTC/CNRS · Paris
Perception Research Engineer at Konboi in Paris, on autonomous trucks and AI retrofit for freight. I build deep learning systems for 3D perception — geometric learning, multi-modal fusion, and models that report when they are unreliable.
Before Konboi I worked on perception at Motional and Toyota Motor Europe. My PhD at Heudiasyc (UTC/CNRS) led to three conference papers and two international patents on camera–LiDAR calibration.
Much of my published work is online extrinsic calibration: estimating 6-DoF transforms from ordinary driving data, without targets, with uncertainty — PseudoCal, UniCal, MULi-Ev, and WACV 2025 on KITTI, DSEC, and nuScenes.
Selected papers — WACV, CVPR Workshop, BMVC Oral, and related work.
Uncertainty quantification for online calibration via conformal prediction (KITTI, DSEC).
Online deep learning calibration of LiDAR with event cameras.
Single-branch transformer with early multi-modal fusion for 6-DoF alignment. Patent WO2024182787.
Target-free self-supervised geometric learning via pseudo-LiDAR in 3D.
UTC / Heudiasyc, 2025. Online multi-sensor calibration and deep learning for autonomous driving.
Recognised for review contributions to the conference.
Perception Research Engineer — autonomous trucks and AI retrofit for freight.
Defended Deep Learning for Multi-Sensor Calibration in Autonomous Driving at UTC.