Valeo, a global tech company, is offering an internship for a Validation Engineer to develop and apply a methodology for evaluating the 'sim-to-real gap' in ADAS perception algorithms using virtual data. The intern will compare performance metrics between synthetic data and internal Valeo datasets, analyze robustness across various contexts, and propose methodological improvements. This role involves immersion in Valeo's simulation and perception environment, hands-on experience with synthetic data generation tools, and an in-depth study of perception evaluation methods.

Must Have

• Bac +5 level, Engineering school or Master's degree in Artificial Intelligence, Computer Science, Robotics or Computer Vision
• Good proficiency in AI/Deep Learning (PyTorch or TensorFlow)
• Skills in image processing and computer vision (object detection, segmentation, classification)
• Knowledge of ADAS / AD systems and multi-modal perception issues
• Proficiency in Python and data science libraries
• Autonomy, analytical mind, scientific rigor
• Ability to work in a multidisciplinary team
• Fluent English

Good to Have

• Knowledge of synthetic data or simulation pipelines (CarMaker, Carla, PD Replica…)
• Interest in large-scale validation and simulation issues
• Scientific curiosity and an appetite for analyzing technological architectures

Perks & Benefits

• Presence of a social & economic committee (CSE): trips, social and cultural activities
• 70% reimbursement of transport costs

Internship - Validation Engineer: Methodology for virtual data evaluation of the Performance of ADAS Perception Algorithms (M/F)

Join the technological revolution of autonomous and connected vehicles within Valeo!

World leader in driving assistance, Valeo reinvents life on board vehicles and revolutionizes the driving experience. From connectivity and interior control solutions to detection systems, the Valeo Brain division of the Group aims to make mobility safer and more intuitive.

The Valeo Mobility Tech Center based in Créteil (94) is the headquarters of the Valeo Brain Division. This Research & Development center brings together 1180 employees around the design of advanced driving assistance systems (ADAS), connectivity solutions, human-machine interfaces, and cabin monitoring systems.

The Driving Assistance Research Center at Valeo is composed of a multidisciplinary team with strong expertise in upstream technologies related to driving automation (ADAS and Automated Driving). The team contributes to the definition and feasibility studies of innovative systems, as well as their validation through simulation and prototyping.

In this context, you will join the System Design & Virtual Validation (Simulation) team with the main mission of developing and applying a methodology for evaluating the “sim-to-real gap”, in order to compare the performance metrics of perception algorithms (LRCV, BEV, etc.) developed within the department between:

• Scenes from a synthetic data generation platform,
• Scenes from Valeo's internal datasets (baseline).

The internship will take place in several stages:

• Immersion in the Valeo working environment (simulation, perception, internal tools).
• Familiarization with synthetic data generation tools and understanding of associated pipelines, as well as their technological foundations.
• State of the art of global perception evaluation methods (mAP, PR-AUC, F1-score, etc.) applied to multiple classes (lane, object, traffic sign).
• Definition of metrics and implementation of adapted evaluation protocols.
• “Sim-to-real gap” comparison between Valeo data and synthetic data.
• Analysis of the robustness of results according to different contexts (traffic density, brightness, scene complexity, etc.).
• Proposal of methodological improvements or recommendations for the use of synthetic data in the Valeo validation cycle.
• Writing a summary report and presenting the results.

You will be accompanied by our Simulation & Perception engineers throughout the internship. > www.valeo.com

About you:

• Level: Bac +5, Degree prepared: Engineering school or Master's in Artificial Intelligence, Computer Science, Robotics or Computer Vision.
• Good proficiency in AI/Deep Learning (PyTorch or TensorFlow).
• Skills in image processing and computer vision (object detection, segmentation, classification).
• Knowledge of ADAS / AD systems and multi-modal perception issues.
• Proficiency in Python and data science libraries.
• Autonomy, analytical mind, scientific rigor.
• Ability to work in a multidisciplinary team.
• Fluent English.

Additional assets:

• Knowledge of synthetic data or simulation pipelines (CarMaker, Carla, PD Replica…).
• Interest in large-scale validation and simulation issues.
• Scientific curiosity and an appetite for analyzing technological architectures.

Our advantages:

• Presence of a social & economic committee (CSE): trips, social and cultural activities
• 70% reimbursement of transport costs

Why Valeo?

• To join a technological and industrial leader, a French pioneer in automotive innovation
• For a dynamic career with opportunities for permanent contracts, theses or International Corporate Volunteering (VIE Program)
• To contribute to the development of cleaner, safer and smarter mobility

Valeo attaches great importance to diversity, whether cultural, intergenerational, gender or concerning people with disabilities.

Job: R&D Trainee/Apprentice/VIE

Organization: System

Schedule: Full time

Employee Status: Trainee (Fixed Term) (Trainee)

Job Type: Trainee paid

Job Posting Date: 2025-12-01