Research Engineer, SLAM & Multi-View Geometry

📋 Description

• • As a pivotal member of OpenAI's innovative Robotics team, you will be instrumental in advancing the frontiers of general-purpose robotics and pushing towards Artificial General Intelligence (AGI) within dynamic, real-world environments. This role offers a unique opportunity to work across the entire AI model stack, integrating state-of-the-art hardware and sophisticated software solutions to explore a diverse array of robotic form factors. Your contributions will be crucial in seamlessly merging high-level AI capabilities with the inherent constraints of physical systems, ultimately aiming to enhance and improve people's lives.
• • In this specialized role, you will focus on developing and deploying advanced systems that empower robots to perceive, meticulously track, and accurately reconstruct their surroundings in three-dimensional space. This will be achieved through the sophisticated processing of multi-camera and multimodal sensor data. You will be deeply involved in the creation and refinement of real-time and offline Simultaneous Localization and Mapping (SLAM) pipelines. These pipelines are essential for critical operations such as teleoperation and the collection of vast amounts of robot-generated data. Furthermore, you will architect and build scalable systems designed for the comprehensive reconstruction and precise tracking of 3D structures from extensive datasets, paving the way for novel approaches in world modeling and robotic perception.
• • We are actively seeking individuals who possess a robust foundation in computer vision principles, coupled with hands-on, practical experience in constructing resilient and high-performing perception systems. The ideal candidate will demonstrate comfort and expertise in navigating both classical geometry-based methodologies and contemporary machine learning techniques. A collaborative spirit and a passion for working closely with AI researchers and fellow engineers are highly valued, as this role thrives on interdisciplinary teamwork.
• • Your responsibilities will encompass the end-to-end development and deployment of online SLAM systems. These systems will be integral to robotic data collection processes, leveraging advanced multi-camera sensor configurations and sophisticated teleoperation platforms. You will also be tasked with building robust systems for large-scale 3D reconstruction and point tracking, capable of handling massive datasets. This work will directly enable groundbreaking advancements in world modeling and perception capabilities.
• • A significant aspect of this role involves collaborating closely with research and engineering teams to effectively scale multi-view geometry pipelines. This scaling effort is critical for processing the ever-increasing volume of data generated by our robotic systems. You will be dedicated to enhancing the accuracy, reliability, and overall scalability of the perception systems that are fundamental to our robotic data collection and AI training pipelines.
• • Furthermore, you will foster strong collaborations across our robotics, perception, and machine learning teams. This cross-functional synergy will be essential for integrating cutting-edge geometry-based methods with advanced learned models, creating more intelligent and capable robotic agents.
• • This position is based in our vibrant San Francisco, CA office. We operate under a hybrid work model, requiring four days per week in the office, and we are pleased to offer relocation assistance to support new employees joining our team.
• • You will contribute to the core technology that drives our robotic initiatives, directly impacting the development of AI systems that can understand and interact with the physical world. This includes optimizing algorithms for real-time performance, ensuring robustness in diverse and challenging environments, and contributing to the foundational understanding of how robots perceive and navigate complex spaces.
• • The role demands a proactive approach to problem-solving, an ability to translate complex research concepts into practical engineering solutions, and a commitment to pushing the boundaries of what is currently possible in robotics and AI perception.
• • You will be at the forefront of developing the next generation of robotic intelligence, working on projects that have the potential for significant real-world impact. This includes contributing to systems that can learn from experience, adapt to new situations, and perform tasks that were previously considered beyond the reach of automated systems.
• • Your work will directly influence the development of AI models that are trained on real-world data, ensuring that our systems are grounded in physical reality and can operate effectively outside of simulated environments. This is a unique opportunity to shape the future of AI and robotics.