Member of Technical Staff - Pretraining
Black Forest Labs7 months ago
Freiburg im Breisgau, GermanySenior / Staff+
Responsibilities
- Lead large-scale pretraining experiments for multimodal foundation models.
- Develop and evaluate novel ideas across architecture, optimizers, and training algorithms.
- Contribute to low-level GPU optimizations and high-level model design.
- Independently lead focused research projects and drive cross-team initiatives.
Requirements
- Experience leading or co-owning pretraining for a foundation model that shipped to production.
- Ability to make architectural decisions that impact model performance.
- Deep experience with large-scale distributed training and debugging at scale.
- Strong intuition for architecture and objective design.
- Track record of publishing in top venues with production impact.
- Proficiency in Python and PyTorch, with the ability to modify low-level training code.
- Familiarity with visual generative models.
Benefits
- Flexible work arrangements with options for remote work and in-person collaboration.
- Coverage of reasonable travel costs for in-person meetings.
About Black Forest Labs
Black Forest Labs builds generative image and video models for creators, developers, and enterprises, including the FLUX series and work behind Latent Diffusion/Stable Diffusion. It distributes models and tooling via open releases, APIs, and enterprise licensing for image generation and editing workflows. Founded in 2024 and headquartered in Freiburg, Germany, the privately held company focuses on visual AI research with a presence in San Francisco.