Research Engineer - Generative Media
Join a pioneering AI lab to develop frontier generative media experiences. You will design, implement, and evaluate deep learning models for real-time video and audio synthesis, collaborate with research teams, and contribute to datasets and publications.
Responsibilities
- Design, implement, train, and evaluate novel deep learning models for real-time generative media, including video-audio synthesis, manipulation, and understanding.
- Develop and optimize algorithms and systems for AI experiences on Omni platforms, focusing on live dialog, live video editing, and crosscut.
- Collaborate with research scientists and engineers to prototype, experiment, and scale generative AI techniques using software engineering best practices.
- Contribute to the design of datasets, evaluation methodologies, and infrastructure required for training and testing generative models.
- Analyze results to iterate on architectures, contribute to research publications, and help deploy successful research into company products.
Requirements
- Bachelor’s degree or equivalent practical experience.
- 4 years of experience in designing, training, and evaluating deep generative models (e.g., Diffusion Models, Transformers, GANs) for media synthesis and manipulation (image, video, audio).
- 4 years of experience with experimentation, dataset curation, eval systems, and building ML pipelines.
- 4 years of experience with machine learning, computer vision, and deep learning.
Nice to have
- Master's or PhD degree in Computer Science, Machine Learning, Statistics, or a related field, or equivalent practical experience.
- Experience with multimodal learning, integrating video, audio, and text.
- Experience with video generation and editing models.
- Experience with Python and deep learning frameworks such as JAX, TensorFlow, or PyTorch.
- Knowledge of techniques for model optimization and efficient/low-latency inference suitable for real-time applications.
- Familiarity with real-time systems or media streaming technologies.