Senior Research Scientist (Architectures Research)

About Nebius:

Nebius is leading a new era in cloud infrastructure for the global AI economy. We are building a full-stack AI cloud platform that supports developers and enterprises from data and model training through to production deployment, without the cost and complexity of building large in-house AI/ML infrastructure.

Built by engineers, for engineers. From large-scale GPU orchestration to inference optimization, we own the hard problems across compute, storage, networking and applied AI.

Listed on Nasdaq (NBIS) and headquartered in Amsterdam, we have a global footprint with R&D hubs across Europe, the UK, North America and Israel. Our team of 1,500+ includes hundreds of engineers with deep expertise across hardware, software and AI R&D.

The role

Nebius AI R&D conducts frontier applied research to make open-source AI highly competitive for real-world use cases. Our Architectures Research stream explores how models can attend, remember, reason, and adapt more effectively, enabling longer and richer workflows at lower computational cost.

We are looking for a Senior Research Scientist to develop new model architectures and methods in areas such as:

- Efficient, sparse, and adaptive attention

- Long-context models and persistent memory

- Post-training transformation of pretrained models -

- Selective computation and dynamic inference

- New architectures for reasoning and continual adaptation

Responsibilities

- Formulate original research questions and translate them into rigorous experimental programs

- Design and evaluate architectural changes at meaningful model scales

- Develop methods that preserve model quality while reducing training or inference cost

- Collaborate with engineering teams to validate ideas in efficient implementations

- Publish research and contribute to open-source models, methods, and tools

- Mentor researchers and help shape the stream's research direction

What we expect

- A PhD or equivalent research experience in machine learning

- Deep knowledge of transformers, attention, language-model training, and modern model architectures

- A strong publication record or comparable evidence of original research

- Experience designing rigorous experiments and drawing clear conclusions from ambiguous results