Nebius PhD Fellowship Program

Are you a PhD student pushing the frontier of AI? Join the first Nebius PhD Fellowship Program and turn your research into real-world impact.

We support exceptional PhD students whose research aligns with Nebius’s long-term priorities. With funding, mentorship and compute access, you’ll work alongside our research teams to advance your research and help bridge academia and industry.

Funding: Up to $120,000 per award, including the internship
Duration: A 9-month fellowship and a mandatory 3-month internship
Fellowship period: September 2027–May 2028
Application deadline: December 31, 2026

About the fellowship

The Nebius PhD Fellowship supports researchers working on important challenges in inference optimization, post-training and agentic systems.

Through the program, you will receive financial support for your research, mentorship from Nebius researchers, access to compute and hardware, and the opportunity to put your ideas into practice through a 3-month internship with our research team.

The internship can take place before or after the fellowship academic year, on-site in the Bay Area, or in Amsterdam.

Research areas

We welcome proposals aligned with one of the following areas.

Inference Optimization

Help make large-model serving faster, cheaper and more scalable on modern GPU clusters.

Topics include speculative decoding, KV-cache compression and management, low-precision inference such as FP8/FP4 quantization, sparse and long-context attention, disaggregated prefill/decode architectures, MoE serving, and kernel- or scheduler-level optimizations.

Strong proposals combine algorithmic ideas with real systems benchmarks, measuring latency, throughput and cost per token at scale.

Post-Training

Help turn pretrained models into capable, aligned and reliable systems.

Topics include reinforcement learning from human or AI feedback, RL with verifiable rewards, reward modeling, preference optimization, reasoning and chain-of-thought training, synthetic data generation, distillation, and rigorous evaluation of post-trained models.

We especially value work that improves sample efficiency or makes post-training pipelines more reproducible and open.

Agentic Systems

Advance LLM-based agents that plan, use tools and act over long horizons.

Topics include multi-agent orchestration, tool-use and code-execution agents, memory and long-horizon planning, RL environments and training for agents, agent benchmarks and evaluation, and the reliability and safety of autonomous systems.

Proposals should target measurable improvements on realistic tasks, such as software engineering or research automation, beyond demonstrations alone.

Who can apply

You should:

• Be enrolled full-time in a PhD program at an accredited university and remain enrolled throughout the fellowship.

• Have a strong research, publication and conference record.

• Conduct research relevant to one of the program’s three research areas.

• Be available for the mandatory 3-month internship before or after the fellowship academic year.

First-year PhD students are welcome to apply. Your research achievements and publication record matter more than your year of study.

How we evaluate applications

Applications are reviewed by an internal research committee. We consider:

• Research excellence, including your publication and conference record.

• Alignment with Nebius’s research priorities.

• The quality of your research proposal.

• Your portfolio of projects.

• Your recommendation letters.

We will review all applications after the submission deadline of December 31, 2026. Applicants can expect to hear from us after this date, with shortlisted candidates invited to a final interview with the team in January. 

How to apply

Please submit your application by December 31, 2026, including:

• CV.

• Research proposal.

• Portfolio of projects, with links to relevant code, papers or demos.

• Three recommendation letters, including one from your thesis advisor.

• Confirmation of your availability for the mandatory 3-month internship.

You may also include an optional statement of purpose or cover letter.

Preparing your application

Research proposal: Describe your PhD research and post-graduation goals. Explain which research area your proposal aligns with, why it matters to Nebius, and what you aim to achieve. Include your results, publications and conference papers to date, and define concrete ways to measure success.

Portfolio: Select a few of your strongest projects and include relevant links. Prioritize quality and relevance over quantity.

Recommendation letters: Choose people who know your work well, including your thesis advisor. Ask at least two weeks in advance and ensure all three letters arrive before the application deadline.

Optional statement of purpose: Tell us why you’re interested in the fellowship and share anything your CV does not capture.

Program timeline

• November 1, 2026: Applications open.

• December 31, 2026: Applications close.

• January 2027: Final interviews with shortlisted applicants.

• End of January 2027: Fellows announced.

• September 2027–May 2028: Fellowship period.

• Before or after the fellowship academic year: Mandatory 3-month internship.

Additional program information

Can I combine the fellowship with part-time work?
Part-time work may be considered case by case, depending on your profile.

Can I hold another fellowship at the same time?
No. The Nebius PhD Fellowship cannot be held concurrently with another fellowship.

Can I receive the fellowship more than once?
No. This is a one-time fellowship, and previous recipients cannot reapply.

Interested?

Apply to the Nebius PhD Fellowship Program and help advance the next generation of AI research.

Questions? Contact earlytalent@nebius.com