Applied Scientist / Applied Scientistess

Overview

The Spatial AI Lab is a pioneering entity within the Applied Sciences Group at Microsoft, dedicated to the development of next-generation human-computer interaction technologies. Our mission is at the forefront of integrating advanced AI innovations with novel hardware capabilities and diverse device form-factors. Comprising a team of skilled scientists and engineers, our expertise encompasses computer vision and multi-modal AI, with a strong emphasis on spatial and embodied AI.

As a Scientist specializing in Multimodal Foundation Models & Robotics, you will engage in cutting-edge research that intersects large-scale generative modeling with embodied AI, focusing particularly on robotics. Your role will primarily revolve around architecting the core intelligence for the next generation of intelligent agents, training multimodal foundation models to enable these agents to perceive complex environments, reason about various tasks, and operate seamlessly in both physical and digital realms. This is a remarkable opportunity to deepen your expertise in training embodied foundation models and deploying algorithms on robotic hardware and large-scale AI systems. You will also contribute to our pioneering research through publications and collaborations with esteemed partners like ETH Zurich.

At Microsoft, we strive to empower every individual and organization across the globe to achieve more. We foster a collaborative environment where employees embody a growth mindset, drive innovation to empower others, and work together towards shared objectives. Each day, we commit to our values of respect, integrity, and accountability, cultivating a culture of inclusion that enables everyone to thrive in their careers and beyond.

Responsibilities

• Design and implement cutting-edge foundation models and algorithms for general-purpose embodied agents.
• Develop high-performance machine-learning pipelines, optimizing data and learning frameworks for scalability, efficiency, and performance.
• Optimize and deploy AI models on robotic hardware.
• Collaborate with Microsoft research and engineering teams to transition groundbreaking research into real-world applications.
• Drive rapid team progress, with success reflected in both the enhancement of internal capabilities and impactful contributions to the scientific community.
• Engage in collaborations with leading academic partners such as ETH Zurich, contributing to pioneering research at scale, with opportunities to co-author work in prestigious venues, present at workshops, and mentor students.

Qualifications

Preferred Qualifications:

• A PhD in Robotics, Machine Learning, or a related field, or a minimum of 3 years of relevant industry experience.
• Proven skills in coding, debugging, and engineering with programming languages such as Python or C++.
• Hands-on experience with contemporary deep learning frameworks (e.g., PyTorch, TensorFlow, Jax).
• A self-motivated team player with strong problem-solving skills and a desire to learn.
• The ability to effectively communicate complex technical concepts to diverse audiences.

Other Requirements:

• Candidates must meet Microsoft, customer, and government security screening criteria. This includes but is not limited to the following specialized security checks:
• Microsoft Cloud Background Check: Successful completion is required upon hiring/transfer and biannually thereafter.

Additional Qualifications:

• Experience in one or more of the following areas:
• Foundation Models: Hands-on training experience with topics such as LLMs, Large Vision-Language Models (VLMs), video generative models, and action-based transformers.
• Large-Scale ML Systems: Familiarity with extensive machine learning compute systems.
• Robotics: Practical experience in robot learning techniques like reinforcement learning and imitation learning, along with classical control methods.
• A solid understanding of robot kinematics, dynamics, and sensors, as well as familiarity with control algorithms such as PID, model predictive control (MPC), and whole-body control.
• A track record of impact demonstrated through first-author research publications at top-tier conferences in machine learning or robotics (e.g., CoRL, RSS, NeurIPS, ICML, ICLR, CVPR) or successful contributions to industry initiatives.

Application Process

Apply online using the form below. Only applications matching the job profile will be considered.

Microsoft is an equal opportunity employer. All qualified applicants will be considered for employment without regard to age, ancestry, citizenship, color, family or medical care leave, gender identity or expression, genetic information, immigration status, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran or military status, race, ethnicity, religion, sex (including pregnancy), sexual orientation, or any other characteristic protected by applicable local laws, regulations, and ordinances. If you require assistance with religious accommodations or a reasonable accommodation due to a disability during the application process, please read more about requesting accommodations.