World Model Research & Training

At Universal Robots, part of Teradyne Inc., Our vision is to create a world where people work with robots, not like robots. And as the market leader with 75,000+ collaborative robots (cobots) already installed worldwide, were well on the way to achieving it. We employ 1000+ people in offices across North America, South America, Europe and Asia and were growing all the time. Our team is made up smart, creative people working at the forefront of automation. Together we find innovative solutions to some of the most important manufacturing issues facing businesses today. We dare to do what others find impossible- working with advanced technologies to change in the way businesses operate, so if youre looking to build your career with a ground-breaking technology company in dynamic environment with career advancement UR is the place for you.

Our Purpose

At Teradyne Robotics, including Universal Robots (UR) and Mobile Industrial Robots (MiR), our mission is simple:Automation for Anyone. Anywhere.

Our vision is to create a world where people work with robots, not like robots. With more than 100,000 collaborative robots deployed worldwide, Universal Robots leads the global market in collaborative automation while MiR provides cutting-edge autonomous mobile robots that transform internal logistics.

Together we enable manufacturers of all sizes to automate, improve productivity, and empower people through robotics.

Opportunity Overview

• Design, train, and iterate on world models that accurately capture real-world physical dynamics, robot-environment interactions, and long-horizon scene evolution
• Develop scalable training pipelines for world models using large-scale real and synthetic datasets, including data collected from deployed robot fleets
• Research and implement novel architectures — including diffusion-based, autoregressive, and latent space models — to improve world model fidelity, generalisability, and inference efficiency
• Investigate and improve model conditioning strategies that allow world models to generalise across diverse environments, robot embodiments, and task types
• Drive experiments to push the frontier of predictive world modelling, including multi-step rollouts, uncertainty estimation, and sim-to-real transfer

3D scene generation

• Research and develop methods for generating high-fidelity 3D scenes — including geometry, texture, lighting, and object placement — for use in robotics simulation environments
• Explore and implement neural scene representations such as NeRF, Gaussian splatting, and implicit surface models for fast, photorealistic simulation asset generation
• Build pipelines that automatically generate diverse, physically plausible simulation environments from minimal inputs such as floor plans, semantic maps, or real-world scans
• Develop techniques for procedural and generative scene variation to support large-scale data augmentation and robot training at scale
• Ensure generated 3D scenes meet the physical accuracy requirements necessary for meaningful robot policy training and evaluation

• Design and curate training datasets combining real-world robot data, synthetic simulation data, and internet-scale visual data to improve world model and scene generation quality
• Develop data collection, annotation, and preprocessing pipelines that scale with fleet size and simulation throughput
• Work with the engineering team to productionise training workflows on large GPU clusters, ensuring efficient utilisation, reproducibility, and experiment tracking
• Define and implement data quality standards and filtering strategies to maximise signal and minimise distribution mismatch between simulation and real deployments

• Define rigorous evaluation frameworks and benchmarks for world model fidelity, scene generation quality, and downstream robot task performance
• Measure and actively close the sim-to-real gap, ensuring that policies and behaviours trained in simulation transfer reliably to real robot deployments
• Establish KPIs that connect research outputs to product-level goals such as integration time reduction, fleet performance, and scenario coverage
• Publish and present research findings internally and at top-tier venues (CoRL, ICRA, NeurIPS, CVPR) to advance the field and attract research talent

• Work closely with the simulation engineering team to ensure research outputs are productionised effectively into the simulation platform
• Collaborate with robotics engineers and application teams to identify the highest-value research problems grounded in real deployment challenges
• Mentor junior researchers and engineers, raising the overall research capability of the team
• Contribute to shaping the long-term research agenda for world models and 3D generation within the simulations division
• Stay at the forefront of developments in generative AI, 3D vision, and embodied intelligence, synthesising advances into actionable research directions

At Universal Robots, we ask a lot from our employees and in exchange, we offer an exciting job in a dynamic and fast-growing company with the opportunity of working alongside skilled and talented colleagues from many different cultures-

You will be empowered to help yourself and your team become a success by taking responsibility for your own work and the products you help create.

At Universal Robots, we trust our employees to make educated choices that benefit the organization.

In Addition, We Offer

• A collaborative and inclusive work environment
• Opportunities for professional and personal development
• Exposure to cutting-edge robotics technology
• Flexible working conditions
• Health and wellness benefits
• And much more

  We are only considering candidates local to position location and are unable to provide relocation for this position

• This position is not eligible for visa sponsorship
• Universal Robots is an equal-opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.