Physics Informed AI/ML Researcher

Job Requisition ID #

25WD90902

Position Overview

The Computational Sciences group within Autodesk Research is seeking exceptionally talented and experienced Research Scientists to join our team. We are focused on advancing the state-of-the-art for design tools to allow architects, artists, designers, and engineers to Make Anything. Our research centers around generative methods for the design, manufacturing and construction of objects in the built environment and spans physics-based modeling, simulation, design optimization and inverse problems. You will join a group shaping the future of physics informed AI and its applications in areas of interest to Autodesk. You will work closely with experts in geometric modeling, simulation, robotics, high-performance computing, machine learning, sensing, computer vision, industrial manufacturing and construction.

The Physics Informed AI group will be specifically focused on researching and implementing novel algorithms for physics-informed AI for use in design, simulation and engineering tools. We are interested in physics-based generative models, surrogate and reduced-order models, data assimilation and digital twins, physics-based reasoning and the interpretation of physical results using AI.

Responsibilities

• Research and implement physics-informed AI algorithms, and demonstrate their applicability to areas of interest for Autodesk
• Design and build physics-informed AI workflows and data pipelines, create and curate training data, develop and test models
• Work closely with a multi-disciplinary team of research scientists and engineers to conceive, plan, develop, and implement scientific research projects
• Supervise researchers, engineers and interns (undergraduate, graduate or post-doctoral)
• Participate in research collaborations with external research institutes and universities
• Publish papers in peer-reviewed scientific journals/conferences
• • Document and communicate the intent and the results of projects in clear terms to both technical and non-technical team audiences both internally and externally
• Promote the use of research findings throughout Autodesk
• Consult with product teams on the implementation of research into products

• Advanced degree in Computer Science, Engineering, Physics or Applied Mathematics
• Excellent knowledge of computational physics and numerical methods, e.g. for deriving and solving the large linear and non-linear systems that arise from discretizing differential equations
• Knowledge of computational geometry, topology and geometric deep learning
• Extensive experience with methods that combine modern machine learning and AI methods with computational physics algorithms.
• Familiarity with best practices in data creation, data and workflow management as related to training AI models
• Experience creating and training machine learning models at scale (e.g. using PyTorch, TensorFlow, Flax, etc)
• Ability to articulate how physics-informed AI methods relate to more traditional numerical methods, surrogate models, reduced order models and simulation approaches in, for example, CFD or computational mechanics.
• Python and C++. Familiarity with HPC techniques. Familiarity with collaborative development environments and version control systems
• Excellent written and oral communication skills
• Strong publication history in relevant conferences and journals
• Ability to quickly learn new technologies and adapt to new situations
• Ability to collaborate effectively with a diverse, multicultural, global team of scientists, engineers and architects