Machine Learning Scientist 5

The Studio Media Algorithms team is at the forefront of innovation to enhance and support the vision of the creators of movies, TV shows and other multimedia work. This team's work is responsible for increasing member value, and driving efficiency of the content creation process, ultimately creating more joy for viewers all over the world. To learn more about the domain, here are some links related to what we do: Creating Media with Machine Learning and Computer Vision Research at Netflix.

We are looking for a Research Scientist with demonstrated experience in computer vision (CV) and/or related areas such as natural language processing (NLP) or computer graphics (CG) to research and develop core algorithms and models that will be incorporated into the tools used by content creators throughout the production lifecycle, including live action, animation, and games.

In this role, you will

• Design, train/post-train/fine-tune, and evaluate foundational algorithmic solutions with applications in a variety of domains in the VFX, animation, and games space, pushing forward the state of the art in CV as needed
• Develop reusable foundational components and best practices, such as dataset curation, large-scale training, and post-training processes, which can be used by other scientists and engineers across teams
• Work cross-functionally with engineers, scientists, artists, and product leaders to help identify and prioritize strategic research investment opportunities and problem requirements
• Expand technical depth and domain expertise into new/adjacent areas as the business needs and the state of technology evolve
• About you:
• Research experience with a successful track record of delivering quality results and/or academic publications in top ML, CV, CG, or NLP venues
• Deep familiarity with modern generative model architectures, including diffusion models (e.g., MMDiT) and/or autoregressive models (e.g., GPT-style).
• Experience with the full stack of model development: data curation, annotation/captioning, distributed training, post-training techniques (e.g., fine-tuning, RL), and/or robust evaluation.
• Expertise in designing and training deep learning (DL) architectures for media understanding and generation, with a broad understanding of DL methods and literature
• Deep mathematical skills with knowledge of statistical methods and optimization
• Extensive experience with DL frameworks such as Tensorflow or PyTorch
• Strong programming skills in languages such as Python, Java, or C++
• Great interpersonal skills for collaboration with technical and non-technical partners
• Advanced degree in Computer Science or related field
• Bonus experience:
• Cutting-edge work on modern generative methods and their building blocks
• Expertise in 3D vision and/or graphics
• Domain experience in content production, such as live action, games, or animation, and a track record of developing tools/solutions in these (or closely relevant) areas