Intern, World Models for XR and Robotics
InterDigital is
·Today
- Location
- Rennes_France
- Type
- Internship
- Department
- Engineering
- Seniority
- Internship
- Education
- Master
- Source
- Workday
Description
World Models for XR and Robotics
About InterDigital
InterDigital is a global research and development company focused primarily on wireless, video, artificial intelligence (“AI”), and related technologies. We design and develop foundational technologies that enable connected, immersive experiences in a broad range of communications and entertainment products and services. We license our innovations worldwide to companies providing such products and services, including makers of wireless communications devices, consumer electronics, IoT devices, cars and other motor vehicles, and providers of cloud-based services such as video streaming. As a leader in wireless technology, our engineers have designed and developed a wide range of innovations that are used in wireless products and networks, from the earliest digital cellular systems to 5G and today’s most advanced Wi-Fi technologies. We are also a leader in video processing and video encoding/decoding technology, with a significant AI research effort that intersects with both wireless and video technologies. Founded in 1972, InterDigital is listed on Nasdaq.
InterDigital is a registered trademark of InterDigital, Inc.
For more information, visit: www.interdigital.com.
Summary
World models learn an internal representation of an environment and predict how it changes in response to user or agent actions. This six-month internship will investigate their potential for XR (Extended Reality) and robotic interaction, combining a state-of-the-art review with one focused proof of concept (PoC).
The review will cover interactive visual world models, such as Tencent HY-World 1.5 / WorldPlay and Skywork Matrix-Game 3.0, alongside action-conditioned world models for robotics, such as NVIDIA Cosmos. These are starting points for the study, not a fixed shortlist. Visual plausibility and temporal consistency will be distinguished from accurate prediction of physical motion, contact and action outcomes.
For XR glasses or headsets, key questions include advantages of World models integration regarding responsiveness to user movement, spatial consistency compared to existing solutions. For robotics, the study will examine how world models predict the effects of actions and support their evaluation. An Universal Robots, either physical (e.g. UR3e arm) or simulated in Gazebo, will provide a possible experimental platform. An XR interface could display the observed workspace and predicted robot or object motion.
The internship has two phases. The first phase will select a suitable world model and define its integration into the existing Unity and Meta Quest demo (that also includes a virtual robot), while assessing a possible XR and robotics extension.
Phase 1 — State of the art and feasibility assessment. Review world models for both XR and robotics, including their representations, action inputs, prediction horizons, data requirements and evaluation methods. Compare the candidate models on inference latency, GPU and memory needs, availability of code and pretrained weights. Test shortlisted approaches on available hardware and assess integration with the existing Unity demo and Meta Quest headset. Conclude with a model comparison, an integration plan and a feasibility assessment for a possible XR and robotics PoC.
Phase 2 — Development and evaluation. First, integrate the selected world model into the existing Unity demo running with a Meta Quest headset. Adapt existing model code, connect XR tracking and interaction inputs, and display model outputs in Unity. Evaluate responsiveness, spatial and temporal consistency against the existing demo as a baseline. If time, feasibility and equipment availability permit, extend this integration into a PoC assessing the value of world models for a combined XR and robotics experiment, using a physical UR3e or a UR3e simulated in Gazebo (or other environment). For example, display predicted robot or object motion in XR and compare predictions with observed outcomes in a simple manipulation task.
You will work in collaboration with experts from InterDigital located in Rennes, France.
You will have access to AI Tools to assist you in your work.
Responsibilities
- Review world models for interactive XR and action-conditioned robotic prediction
- Compare candidate models and define an integration plan for the existing Unity and Meta Quest demo
- Deploy and benchmark shortlisted models on available hardware
- Connect XR tracking and interaction inputs to the world model and display its outputs in Unity
- Integrate the world model into the existing Meta Quest demo; if feasible, develop an XR and robotics PoC
- Measure latency, spatial and temporal consistency, compute load and XR experience quality
- Assess added value against the existing demo; for an optional robotics PoC, evaluate prediction error and task outcomes
- Deliver reproducible experiments, documented code and a final presentation
Qualifications
- Final-year M.Sc. or engineering student in Computer Science, Machine Learning, Robotics, Signal/Image Processing or a related field
- Good skills in Python and PyTorch
- Experience with generative models (diffusion, autoregressive transformers, VAE) is a plus
- Interest in XR, robotics, world modeling and interactive systems
- Experience with Unity and XR application development for Meta Quest headsets
- Experience with the UR3e robot arm and Gazebo simulation is a plus; familiarity with ROS 2, kinematics or robot control is also an asset
- Knowledge of computer vision, 3D modeling, networking or video coding is a plus
- Ability to read research papers with a critical eye
- Experience with AI Tools for software development is a plus
- Professional English proficiency (French is a plus)
- Good communication skills, team player
Keywords
World Models, Generative AI, XR, Unity, Meta Quest, Robotics, UR3e, Gazebo, Action Prediction, Interactive Systems, Modeling and Simulation, PyTorch
Expected Outcomes
- State-of-the-art report covering XR and robotics, with a model comparison, a Unity and Meta Quest integration plan and an assessment of a possible robotics extension
- A working world-model integration in the existing Unity demo using a Meta Quest headset, with documented code, setup instructions and a demonstration
- If feasible, an additional PoC demonstrating and assessing the value of world models for a combined XR and robotics experiment
- Evaluation report comparing the integrated demo with the baseline, covering latency, spatial and temporal consistency, visual quality and resource use; for an optional robotics PoC, prediction error and task outcomes
- Final presentation with findings, limitations and recommendations for future research
- Potential invention opportunities
Location: Rennes, France
Mentors : [Thierry FILOCHE, Bertrand LEROY, Stéphane ONNO]
InterDigital is an equal employment opportunity employer. InterDigital will not engage in or tolerate unlawful discrimination with regard to any employment decision, policy or practice based on a person’s sex, gender, pregnancy (including childbirth, breastfeeding and related medical conditions), age, race, color, religion, creed, national origin, ancestry, citizenship, military status, veteran status, mental or physical disability, medical condition, genetic information, sexual orientation, gender identity or expression, or any other factor protected by applicable federal, state or local law. This policy applies to all terms and conditions of employment, including, but not limited to, recruiting, hiring, compensation, benefits, training, assignments, evaluations, coaching, promotion, discipline, discharge and layoff.