Research Scientist (Generative Modeling)

What You Will Do

• Design, implement, and train large-scale diffusion models for generating 3D worlds
• Develop and experiment with large-scale diffusion models to add novel control signals, adapt to target aesthetic preferences, or distill for efficient inference
• Collaborate closely with research and product teams to understand and translate product requirements into effective technical roadmaps.
• Contribute hands-on to all stages of model development including data curation, experimentation, evaluation, and deployment.
• Continuously explore and integrate cutting-edge research in diffusion and generative AI more broadly
• Act as a key technical resource within the team, mentoring colleagues, and driving best practices in generative modeling and ML engineering

Key Qualifications

• 3+ years of experience in generative modeling or applied ML roles, ideally at a startup or other fast-paced research environment
• Extensive experience with machine learning frameworks such as PyTorch or TensorFlow, especially in the context of diffusion models and other generative models
• Deep expertise in at least one area of generative modeling: pre-training, post-training, diffusion distillation, fine-tuning with new conditioning signals, etc for diffusion models
• Strong history of publications or open-source contributions involving large-scale diffusion models
• Strong coding proficiency in Python and experience with GPU-accelerated computing.
• Ability to engage effectively with researchers and cross-functional teams, clearly translating complex technical ideas into actionable tasks and outcomes.
• Comfortable operating within a dynamic startup environment with high levels of ambiguity, ownership, and innovation.

Preferred Qualifications

• Contributions to open-source projects in the fields of computer vision, graphics, or ML.
• Familiarity with large-scale training infrastructure (e.g., multi-node GPU clusters, distributed training environments).
• Experience integrating machine learning models into production environments.
• Led or been involved with the development or training of large-scale, state-of-the-art generative models

While not required, experience in one or more of the following areas is a strong plus:

• Large-scale model training
• Data curation for pretraining or post-training
• Tokenizers and VAEs for image, video, or 3D data
• Long-context architectures
• 3D vision