Position Overview

The work we do at Autodesk touches nearly every person on the planet. By creating software tools for making buildings, machines, and even the latest movies, we influence and empower some of the most creative people in the world to solve problems that matter. As an AI Research Scientist in the Autodesk Research AI Lab, you will be doing fundamental and applied research that will help our customers imagine, design, and make a better world.

As a member of the AI Lab, you will be working side-by-side with world-class researchers and engineers to build new ML-powered product features that will help our customers imagine, design, and make a better world. The AI Lab is active in the wider research community and has strong collaborations with top academic labs, combining the best of an academic environment with product-guided research. The AI Lab actively publishes at CVPR, NeurIPS, ICML, ICLR, SIGGRAPH, and other top-tier conferences. A list of recent publications can be found here: https://www.research.autodesk.com/publications/#/selected-filter/publications-research-area-ai-lab

You are a research scientist who specializes in machine learning at scale. You have experience individually and as part of a team training large machine learning models with distributed deep learning. Your skills span across the design of large machine learning models, running large scale experiments informed by scaling laws, and optimizing data and compute GPU throughput to perform efficient model training. You are passionate about solving problems and building things. You are excited to collaborate with AI researchers to build state-of-the-art generative AI features in Autodesk products. You are a good communicator and comfortable working at the intersection of research & product.

The location of this role is flexible. We are a global team, located in London, San Francisco, Toronto, and remotely. Autodesk is a flexible hybrid-first company, allowing workers to work remotely, in an office, or a mix of both.

Responsibilities

• Lead the development and scaling of machine learning models at the intersection of research and product

• Leverage distributed deep learning frameworks, such as DeepSpeed, FSDP etc…, to perform scaling experiments with the support of engineering and data teams

• Collaborate with the engineering teams to find and resolve distributed deep learning training bottlenecks such as checkpointing, model/optimizer compression etc.

• Optimize data flows through the model and overall architecture

• Leverage frameworks such as Open Neural Network Exchange (ONNX) and TensorRT to create high performance model endpoints

• Writing robust, testable code that is well documented and easy to understand

• Analyze errors and provide solutions to problems that arise

• Present the results of your work to collaborators and leadership

Minimum Qualifications

• BSc or MSc in Computer Science, or equivalent industry experience

• Proven experience with large scale distributed ML model training

• Experience with 3D Parallelism including Tensor Parallelism, Sequence Parallelism and Pipeline Parallelism

• Knowledge of frameworks used for model sharding, e.g. DeepSpeed, FairScale

• Experience with large scale distributed data processing

• Experience with software version control, unit tests, and deployment pipelines

• Experience with cloud services & architectures (AWS, Azure, etc.)

• Excellent written communication skills to document code, architectures, and experiments

• Experience with hardware level systems for high performance computing

Preferred Qualifications

• Knowledge of the design, manufacturing, AEC, or media & entertainment industries

• Experience with Autodesk or similar products (CAD, CAE, CAM, etc.)

• Strong data modelling, architecture, and processing skills with varied data representations including 2D and 3D geometry

• C++/CUDA experience

The Ideal Candidate

• A self-starter with initiative to search for solutions and execute on problems with minimal supervision

• A curious, creative problem-solver who is excited to learn & develop new technologies

• A fast learner & excellent communicator who can effectively collaborate across global locations

• Comfortable working in newly forming ambiguous areas where learning and adaptability are key skills