Machine Learning Engineer - Robot Manipulation

<h1>Company Overview</h1> <p>Maven Robotics is building the world’s leading general-purpose AI robots.</p> <p>We are currently operating in stealth and are growing the world’s best team in AI robotics. We are looking for self-starters that are the world’s best in their field, who can innovate from a deep understanding of the fundamentals, and who share our values of unwavering truth seeking and integrity, humility, curiosity, and relentless determination.</p> <h1>Role Description</h1> <p>We are looking to recruit an exceptional <strong>Machine Learning Engineer - Robot Manipulation</strong> to design, implement, test, and deploy robot manipulation algorithms that enable assembly and material movement tasks.</p> <p>In this role you will:</p> <ul> <li>Design and implement machine learning algorithms, with a focus on reinforcement learning (RL) and imitation learning (IL), to enable robotic manipulators to perform complex tasks in dynamic environments.</li> <li>Translate high-level objectives into machine learning problems and deploy robust, scalable models to real-world robotic systems.</li> <li>Integrate your ML solutions into existing robotics workflows, ensuring that models are performant in both simulated and real-world settings.</li> <li>Drive innovation by incorporating the latest research in machine learning into practical applications that push the boundaries of robotic manipulation.</li> <li>Take ownership of critical ML projects, seeing them through from conception to successful deployment.</li> <li>Collaborate across disciplines to ensure seamless integration of ML models and provide technical mentorship to junior engineers.</li> </ul> <h1>Qualifications</h1> <p><em>Must-have:</em></p> <ul> <li>MS or PhD in machine learning, computer science, robotics, or a related field.</li> <li>Strong practical experience in training and deploying machine learning models for real-world applications.</li> <li>Deep understanding of reinforcement learning (RL) and imitation learning (IL) and their application to robotics.</li> <li>Proficiency in programming languages and tools commonly used in machine learning (e.g., Python, PyTorch).</li> <li>Experience with data collection, preprocessing, and management in the context of training ML models.</li> <li>Self-starter attitude with strong ability to identify problems, prioritize them, then plan and execute working solutions.</li> <li>Enthusiasm for working in a fast paced startup environment and eagerness to support the team on a variety of topics.</li> </ul> <p><em>Nice-to-have:</em></p> <ul> <li>Familiarity with robotic simulation environments (e.g., Gazebo, MuJoCo) and experience in sim-to-real transfer.</li> <li>Experience in:</li> <ul> <li>Designing and implementing reward functions for complex manipulation tasks.</li> <li>Developing models that can handle noisy, incomplete, or sparse data.</li> <li>Deployment of ML models to edge devices for real-time inference.</li> <li>Accelerating ML training processes using GPU, TPU, or other HW accelerators.</li> <li>Using reinforcement learning frameworks, e.g. Stable Baselines, RLlib, or similar.</li> </ul> <li>General knowledge of robotics principles, including kinematics, dynamics, and control.</li> <li>Publications or contributions to the machine learning community, particularly in areas related to robotics or reinforcement learning.</li> </ul>