Principal Machine Learning Operations Developer for AI Research, Autodesk

• BSc or MSc in Computer Science or related field, or equivalent industry experience
• Experience with distributed systems for machine learning and deep learning at scale
• Strong knowledge of ML infrastructure and model parallelism techniques, including frameworks like PyTorch, Lightning, Megatron, DeepSpeed, and FSDPProficiency in Python and strong software engineering practices
• Experience with cloud services and architectures (AWS, Azure, etc.)
• Familiarity with version control, CI/CD, and deployment pipelines
• Excellent written documentation skills to document code, architectures, and experiments
• A self-starter who can solve problems with minimal supervision while collaborating effectively with a global, remote-first team
• Adaptable and creative, comfortable building new infrastructure or working within existing codebases
• Thrives in ambiguous, rapidly evolving areas where learning and flexibility are essential
• Excellent communicator who can convey complex technical concepts clearly to diverse audiences

Desirable

• As a MLOps Developers at Autodesk Research, you will be working side-by-side with world-class AI researchers to build and scale foundation models trained on design data
• You will focus on overcoming the challenges associated with large-scale model training and processing of vast amounts of diverse design data
• Your expertise in distributed systems, ML infrastructure, and data engineering will be crucial in developing the next generation of ML-powered product features that will help our customers imagine, design, and make a better world
• You'll be joining a rapidly growing team working on a project that aims to revolutionize the design of nearly every aspect of the built environment. Your contributions will directly influence how designers, architects, and engineers interact with AI tools in the future
• Support AI researchers by building scalable ML training pipelines and infrastructure for foundation model development
• Design efficient data processing workflows for large-scale design datasets and industry-specific file formats
• Optimize distributed training systems and develop solutions for model parallelism, checkpointing, and efficient resource management
• Analyze performance bottlenecks and provide solutions to scaling problems
• Implement and maintain robust, testable code that is well documented and easy to understand
• Collaborate on projects at the intersection of research and product with a diverse, global team of researchers and engineers
• Present results to collaborators and leadership