10 months ago
Remote, United Kingdom +3 moreSenior / Staff+
Responsibilities
- Conduct experiments on efficient training of large language models using interaction traces from varied environments.
- Explore guided generation and search methods in trajectory spaces.
- Develop web-scale methods for mining relevant data and using it efficiently in model post-training.
- Experiment with reinforcement learning configurations in verifiable domains.
- Investigate methods for training AI agents on tasks with non-verifiable reward signals.
- Design, execute, and analyze machine learning experiments with statistical rigor.
- Document research findings and contribute to technical publications or reports.
- Collaborate with adjacent teams to apply research findings in practice.
Requirements
- Profound understanding of theoretical machine learning and reinforcement learning foundations.
- Deep expertise in modern deep learning for language processing and generation.
- Substantial experience training large models across multiple computational nodes.
- Strong software engineering skills, primarily using Python.
- Deep experience with modern deep learning frameworks, particularly JAX.
- Experience designing, executing, and statistically analyzing machine learning experiments.
- Ability to formulate research questions, test hypotheses through experiments, and draw meaningful conclusions.
- Ability to document research findings clearly and contribute to technical publications or reports.
- Experience with deep reinforcement learning for large language models, including reward modeling, DPO, or PPO, is preferred.
- Familiarity with RoPE, ZeRO/FSDP, Flash Attention, and quantization is preferred.
- A bachelor’s degree in Computer Science, Artificial Intelligence, Data Science, or a related field is preferred; a master’s degree or PhD is further preferred.
- Experience delivering products in a dynamic startup-like environment is preferred.
- Experience engineering complex distributed data processing systems or high-load web services is preferred.
- Open-source projects demonstrating engineering ability are preferred.
- Strong English-language, writing, articulation, communication, and leadership abilities.
- Proficiency with contemporary software engineering practices, including CI/CD, version control, and unit testing.
Benefits
- Comprehensive benefits package
- Flexible working arrangements
- Professional growth opportunities
- Collaborative work environment
Tech Stack
Categories
AI ResearchML Engineering
About Nebius
Nebius builds a full-stack AI cloud offering GPU compute, storage, and tools for training and deploying ML models for startups, enterprises, and research labs. It sells consumption-based cloud infrastructure (IaaS/PaaS) and managed services tailored to generative AI workloads, including large-scale model training and inference. The company is headquartered in Amsterdam and operates as an independent provider.