Mathieu Cocheteux

Computer Vision Engineer / Researcher

Thesis completion: April 2025

Seeking Full-Time Industry Positions starting April 2025, with preferred locations:

USA Switzerland Singapore Canada France Europe

Open to opportunities worldwide

          
        Multiple publications at top-tier Computer Vision conferences (WACV, CVPR)

        International patent in Camera-LiDAR calibration technology

        Industry experience at Motional and Toyota Motor Europe

About Me

I am a Doctoral Researcher in Computer Vision specializing in sensor fusion, deep learning for perception, and real-time calibration for autonomous driving.

My research focuses on developing robust deep learning methods for multi-sensor perception, with expertise in LiDAR-camera perception and uncertainty estimation. Through my work at Motional and Toyota Motor Europe, I've developed practical solutions that have resulted in publications in top-tier conferences and an international patent.

Technical Expertise

Machine Learning Deep Learning Uncertainty Estimation Computer Vision Perception 3D Vision Object Detection Sensor Fusion Calibration Autonomous Driving Robotics PyTorch Lightning OpenCV Python C++ ROS ONNX TensorRT TensorFlow Git Docker Linux

Research & Publications

WACV 2025

Uncertainty-Aware Online Extrinsic Calibration: A Conformal Prediction Approach

Key contribution: Novel uncertainty quantification approach for sensor calibration that provides reliability guarantees and improves odometry accuracy.

PDF Poster Video

CVPR 2024 Workshop

MULi-Ev: Maintaining Unperturbed LiDAR-Event Calibration

Key contribution: First method for online calibration of LiDAR with event cameras.

PDF Poster

BMVC 2023 (oral)

PseudoCal: Towards Initialization-Free Camera-LiDAR Calibration

Key contribution: First method for deep learning-based sensor calibration without requiring manual initialization, improving operational robustness.

PDF Video

arXiv 2023

UniCal: A Single-Branch Transformer-Based Model for Camera-to-LiDAR Calibration and Validation

Key contribution: Efficient Transformer-based approach for both calibration and validation, with industry applications leading to an international patent (WO/2024/182787).

PDF

Patents

International Patent

WO/2024/182787 - CAMERA-TO-LIDAR CALIBRATION AND VALIDATION MODEL

Filed through Motional, based on the UniCal transformer-based camera-LiDAR calibration research.

View Patent

Professional Experience

For detailed responsibilities and achievements, please refer to my resume.

Doctoral Researcher, Université de technologie de Compiègne & CNRS

10/2021–04/2025, France

Research Engineer Intern, Motional

06/2022–12/2022, Singapore

Research Engineer, Université de technologie de Compiègne & CNRS

04/2021–10/2021, France

Research Engineer Intern, Toyota Motor Europe

07/2020–01/2021, Belgium

Software Engineer Intern, Safran Electronics & Defense

02/2019–07/2019, France

Education

Additional details about my educational background can be found in my resume.

PhD in Computer Vision

Université de technologie de Compiègne (Sorbonne University alliance) & CNRS, France (Expected 2025)

MSc in Computer Science / Diplôme d'Ingénieur

Université de technologie de Compiègne (Grande École), France (2021)

Contact

I'm actively seeking full-time positions starting April 2025. If you're interested in discussing potential opportunities or have questions about my research, I'd love to connect!

Email

JavaScript required to view email

linkedin.com/in/mathieu-cocheteux

Ready to discuss opportunities?

I'm excited to bring my expertise in computer vision, perception systems, and sensor calibration to challenging industry roles.

Download Resume Contact Me