Member Of Technical Staff - VLM (AI)

TL;DR

Member Of Technical Staff - VLM (AI): Developing and integrating state-of-the-art multimodal vision-language models into the FLUX stack with an accent on novel architectures and multimodal understanding. Focus on pretraining VLMs, optimizing cross-modal alignment, and ensuring scalable deployment without compromising generation quality.

Location: Hybrid in Freiburg (Germany) or San Francisco (USA) (at least 2 days/week or one full week every other week), or remote with a required monthly in-person week.

Company

A frontier research lab behind foundational generative technologies including Latent Diffusion, Stable Diffusion, and FLUX.

What you will do

• Lead the development and training of SOTA multimodal vision-language models within the FLUX stack, innovating on architectures.
• Design fine-tuning strategies for specialized creative use cases such as captioning, editing instructions, and prompt enhancement.
• Research integrations between VLM/LLM capabilities and diffusion/flow pipelines to improve generation quality and controllability.
• Evaluate and translate emerging multimodal research into practical production improvements.

Requirements

• Proven experience pretraining or significantly advancing a VLM that was deployed in production or released publicly.
• Strong publication record or an unambiguous production track record in multimodal architectures.
• Deep understanding of tokenization, alignment, grounding, cross-modal attention, and their respective failure modes.
• Experience with multi-node distributed training at scale.
• Ability to operate effectively at the boundary between research and production.

Nice to have

• Experience with diffusion or flow-based generative models and their composition with autoregressive paradigms.

Culture & Benefits

• Guided by core values: Obsessed (research excellence), Low Ego (best idea wins), Bold (ambitious shipping), and Kind (people over politics).
• Distributed team structure that prioritizes high-quality in-person connection.
• Company covers reasonable travel costs for remote employees to attend mandatory monthly in-person weeks.