Postdoctoral Scholar - SAF Lab, Compass

## Job Title Postdoctoral Scholar - SAF Lab, Compass ## Responsibilities - Push forward the fundamental science of safe autonomy from various perspectives: theoretic contributions, integration with learning, or synthesis from perception. - Develop simulation and evaluation pipelines for complex and large-scale validation in high fidelity simulation environments. - Develop sim-to-real transfer pipelines to deploy simulation-based methods (controllers, policies) on hardware. - Deploy methods on hardware, focusing on dynamically stable robots, validate the underlying science, and identify gaps between science and practice to drive innovation. - Publish research at top-tier robotics, control, and ML venues and contribute to Amazon's scientific reputation in advanced robotics. - Collaborate with product teams and science leaders to set a science roadmap with eventual impact on real robots. ## Requirements - PhD in Computer Science, Robotics, Control, Mechanical Engineering, Electrical Engineering, or a related field with a focus on control, learning, and/or robotics. - Deep understanding of safety-critical control, including control barrier functions and safety filters. - Proficiency in C++ and Python with experience implementing control algorithms and/or learning policies. - Experience with physics simulators for robotics (e.g., Isaac Gym/Sim, MuJoCo, PyBullet). - Experience validating on physical robotic hardware (not simulation-only). - Track record of publications at top-tier venues in control and robotics (e.g., RSS, ICRA, IROS, CDC, CoRL, NeurIPS, ICLR, L-CSS, RAL, TRO, TAC). ## Nice to Have - Understanding of locomotion, reduced order models, layered control architectures, nonlinear control, reachability methods, and whole-body control. - Knowledge of learning-based approaches to robotics (e.g., reinforcement learning, diffusion, VLAs, VLMs, world models). - Exposure to learning-based approaches for CBF synthesis (e.g., neural CBFs, data-driven barrier functions) and integration of CBFs into learning (e.g., CBF-RL). - Understanding of control systems engineering with a focus on layered architecture used in robotic systems (high-level planning, mid-level trajectory generation, low-level feedback control). - Experience with perception on robotic systems (e.g., depth camera and LiDAR based sensing modalities, sensor fusion, semantic tagging). - Familiarity with Hamilton-Jacobi reachability analysis and its relationship to CBF-based approaches. - Knowledge of safety-constrained RL (e.g., constrained MDPs, Lagrangian methods, shielding, CBF-based policy filtering). - Experience with model-based control (MPC, whole-body QP controllers, operational space control) and/or simulation-based predictive control (MPPI). - Experience with hierarchical RL, skill composition, distillation, and multi-task policy architectures for locomotion. - Familiarity with real-time deployment constraints (latency budgets, onboard compute limitations, control-loop frequencies). - Experience building or contributing to large-scale RL training infrastructure (distributed training, GPU clusters). - Strong communication skills and ability to work across disciplinary boundaries (ML, controls, mechanical engineering). ## Benefits - Unique opportunity to collaborate with the inventor of CBFs, top scientists and engineers at Amazon, and establish connections with top academic research labs. - Research in the SAF lab will lay the foundations of safe learning on complex robots, removing bottlenecks to deployment and enabling safe operation around humans. ## About the Role The SAF Lab is the first industry research lab in safe autonomy, developing a universal safety layer for the next generation of robotic systems: mobile robots, manipulators, mobile manipulators, and future platforms with dynamic stability. The role involves pushing the frontiers of performant safety for highly dynamic robots, integrating CBF theory with perception and learning, and evaluating on next-generation robots. The work will underpin robots operating alongside people at Amazon's unprecedented scale.