Inside Nebius Research: Priorities, Open Questions and the PhD Fellowship

Where could your PhD research contribute to the next generation of AI systems?

Join Nebius researchers for an inside look at our work across inference optimization, post-training, and agentic systems. You’ll hear about the questions our R&D teams are exploring, the results they’re seeing, and the challenges that remain open.

We’ll also introduce the Nebius PhD Fellowship and explain what support it offers, who can apply, what fellows can expect, and how they work with Nebius researchers.

What you’ll learnWhat you’ll learn

• Inference optimization: Open challenges in speculative decoding, KV-cache management, low-precision inference, MoE serving, kernels, and scheduling.
• Post-training: Approaches to reinforcement learning, reward modeling, preference optimization, reasoning, synthetic data, distillation, and evaluation.
• Agentic systems: How LLM agents are trained and evaluated to plan, use tools, and act reliably over long horizons.
• How Nebius supports PhD researchers: Learn what the Nebius PhD Fellowship program offers and how fellows collaborate with Nebius teams

Who should attendWho should attend

PhD students, as well as professors and research group leaders exploring joint research with Nebius. The webinar is also relevant to researchers working on AI systems, post-training, or agentic AI