Researcher, Robot Intelligence
Job description
About the role
This position advances robotics through research and development of intelligent systems for manipulation, navigation, and complex reasoning. The hire owns the design of autonomy architectures that interpret high-level task instructions and generate low-level task commands to support complex, goal-driven robot behaviors. They own the development of efficient 3D scene graph representations and other 3D semantic and geometric models and solutions that support scalable perception and reasoning for intelligent robots. The role involves owning scalable robot evaluation pipelines across robot fleets that combine simulation and real-world testing to validate performance and reliability at scale. The hire will own the generation of patents and scientific research papers for publication in top-tier robotics, computer vision, and machine learning venues to advance the field and demonstrate impact. They own demonstrating a solid understanding of robotics fundamentals, including locomotion, manipulation, sensor inputs, pose tracking, 3D mapping, SLAM, ROS, and policy architectures for semi-structured and unstructured environments. The position requires owning show experience with foundation models, vision-language-action models, vision language models, and agentic architectures to enable scalable robot learning.
Key facts
What you'll do
Design of autonomy architectures interpreting high-level task instructions and generating low-level task commands to support complex, goal-driven robot behaviors.
Development of efficient 3D scene graph representations and other 3D semantic and geometric models and solutions supporting scalable perception and reasoning for intelligent robots.
Scalable robot evaluation pipelines across robot fleets combining simulation and real-world testing to validate performance and reliability at scale.
Generation of patents and scientific research papers for publication in top-tier robotics, computer vision, and machine learning venues advancing the field and demonstrating impact.
Demonstration of a solid understanding of robotics fundamentals, including locomotion, manipulation, sensor inputs, pose tracking, 3D mapping, SLAM, ROS, and policy architectures for semi-structured and unstructured environments.
Show experience with foundation models, vision-language-action models, vision language models, and agentic architectures to enable scalable robot learning.
Application of solid computer vision techniques such as object detection, segmentation, tracking, and real-time video tokenizer design, along with multi-view image processing.
Use of tools and processes to monitor model performance and data quality, including model tuning experience to ensure reliable deployment.
Programming in PyTorch, TensorFlow, C, C++, Java, ROS, and CUDA to build and optimize robotic systems.
Solving problems independently and working effectively within cross-functional teams.
Communication of results clearly through strong verbal and written skills for technical presentations and documentation.
Execution of research experiments with rigorous methodology, reproducibility, and clear reporting as core standards.
Maintenance of detailed lab books or version-controlled analysis code to track progress and enable collaboration.
Collaboration with multiple groups including Reasoning, Dexterity, Data Efficiency, and Robot Systems to align research with product goals.
Requirements
A PhD in EECS/Robotics or an equivalent combination of education, training, and experience is required.
Bring 2 14+ years of robotics experience, demonstrating technical depth and real-world impact.
Maintain a strong research track record with publications in top-tier venues such as ICRA, RSS, Science Robotics, CVPR, ICCV, ICML, and NeurIPS.
Demonstrate a solid understanding of robotics fundamentals, including locomotion, manipulation, sensor inputs, pose tracking, 3D mapping, SLAM, ROS, and policy architectures for semi-structured and unstructured environments.
Show experience with foundation models, vision-language-action models, vision language models, and agentic architectures to enable scalable robot learning.
Apply solid computer vision techniques such as object detection, segmentation, tracking, and real-time video tokenizer design, along with multi-view image processing.
Use tools and processes to monitor model performance and data quality, including model tuning experience to ensure reliable deployment.
Program in PyTorch, TensorFlow, C, C++, Java, ROS, and CUDA to build and optimize robotic systems.
Solve problems independently and work effectively within cross-functional teams.
Communicate results clearly through strong verbal and written skills for technical presentations and documentation.
Practical notes
This role is based in Mountain View, California.
The position requires office-based work, sitting and standing at a desk, and frequent use of standard office equipment.
Samsung adheres to equal employment opportunity and reasonable accommodation laws.
Typical interview steps
Research interviews usually include a presentation of past work, a technical discussion, and sometimes a research proposal exercise. Candidates may be asked to design a study or critique a method. Depth of understanding is tested more than speed. Interviewers often ask you to present your past work in depth. Being ready to defend every methodological choice is the core preparation.
Good to know
The role focuses on core robotics research, including perception, planning, and control.
Common tools include PyTorch, TensorFlow, ROS, and CUDA for model development and deployment.
Candidates should expect to work with real-world robot platforms and large-scale evaluation benchmarks.
Research outcomes often appear at top conferences and in patents.
The position contributes to Samsung's broader AI and robotics product strategy.
Career growth
Research careers grow from junior researcher to senior scientist, principal, and lab or research director roles. Some people move into applied research and product work. Publication record or demonstrable impact drives progression, depending on the setting. Research careers reward a strong publication or delivery record. Applied research roles value impact on products as much as novelty.