About the job
About World Labs:
At World Labs, we are at the forefront of developing foundational world models that can perceive, generate, reason, and interact with the 3D world. Our mission is to unlock the full potential of AI through spatial intelligence, transforming visual perception into actionable insights and creative output.
We believe that spatial intelligence will pave the way for innovative storytelling, creativity, design, simulation, and immersive experiences in both virtual and physical realms.
Our diverse team is composed of world-class professionals united by a shared passion for technology and a deep commitment to excellence. We blend expertise from AI research, systems engineering, and product design, creating a dynamic feedback loop between groundbreaking research and user-centric products.
Role Overview
We are seeking a dedicated 3D Reconstruction Specialist to innovate and enhance cutting-edge methodologies for reconstructing high-quality 3D geometry and appearance from real-world data. This role will primarily focus on modern reconstruction techniques—both feed-forward and optimization-based—while emphasizing novel representations, robust optimization, and scalable training and inference workflows.
This is an exciting, hands-on position ideal for individuals who thrive at the intersection of computer vision, graphics, and machine learning. You will collaborate closely with research scientists, ML engineers, and product teams to translate advanced reconstruction concepts into production-ready systems that enhance our core product functionalities.
What You Will Do:
- Design and implement state-of-the-art 3D reconstruction systems, utilizing both feed-forward and optimization-based methods for geometry, appearance, and scene comprehension.
- Research, prototype, and productionize advanced 3D representation techniques (e.g., implicit functions, point-based, or volumetric methods, hybrid representations) focusing on accuracy, efficiency, and scalability.
- Develop and refine optimization pipelines for multi-view reconstruction, including camera pose estimation, joint geometry/appearance optimization, and robust loss formulations.
- Create comprehensive training and evaluation workflows for 3D reconstruction models, from data preparation and supervision strategies to large-scale experiments and performance metrics.
- Collaborate with data and infrastructure teams to seamlessly integrate reconstruction methods with existing 3D data pipelines, rendering systems, and downstream applications.
- Analyze failure modes and data quality issues in real-world reconstruction scenarios, and devise principled solutions to enhance robustness and generalization.

