Member of Technical Staff - Pretraining
Black Forest Labs · Freiburg (Germany)
CHECKED 11H AGO
Frontier labonsitestaff$146k–$381k
At a glance
Highlights
- Lead large-scale pretraining for joint image, video and audio foundation models.
- Shape training objectives and scaling strategies for next-wave visual intelligence.
- Take research from experiment to products used by millions of creators worldwide.
What they offer
- Equity
- Reasonable travel costs covered
Ideal candidate
A frontier-minded researcher who has led pretraining end-to-end and thrives owning bold bets. Obsessed with deep science, low-ego and kind, you want your work shaping products used by millions.
Requirements
- Led or co-owned pretraining for a foundation model that shipped to production or major release
- Own architectural calls such as attention patterns, modulation schemes, loss formulations and tokenization
- Deep experience with large-scale distributed training at 500+ GPU scale including debugging
- Track record of top-venue publications paired with production impact or frontier lab wins
- Deep Python and PyTorch proficiency with low-level training code
- Familiarity with visual generative models
- remote policy
- Hybrid in Freiburg or SF, or remote with monthly in-person week
- hiring process
- AI-assisted tools help manage and organize applications; all hiring decisions are made by the team
Lab profile
About Black Forest Labs
Team behind Stable Diffusion and FLUX building frontier image and video generation models.
What they build
- FLUX image generation models
- FLUX video generation models
Known for
- Pioneers of Latent Diffusion and Stable Diffusion
- State-of-the-art image and video generation
- Models used by millions of creators and businesses
Key facts
- Headquartered in Freiburg, Germany
- Presence in San Francisco