At JetBrains, code is our passion. Since 2000, we've focused on reducing routine work so developers can spend more time building and shipping.

The JetBrains Research team studies machine learning techniques and agent-based approaches for software development to create methods that enable the next generation of ML for code features. We're looking for a researcher to develop and improve training methodologies for models capable of solving tasks such as code generation, code editing, and empowering AI agents. You don't need experience in every area, but you should be comfortable working across topics and delivering publishable results.

In this role, you will:

• Train and scale neural networks for processing source code.
• Develop new methods and improve existing ones for code generation, code editing, and agent-based workflows.
• Mentor colleagues on ML topics.

We're looking for someone who:

• Has experience in natural language processing or code modeling.
• Has worked on distributed training and scaling neural networks.
• Has experience with modern training techniques and architectures (for example, MoE).
• Has publications at top ML conferences.
• Has strong programming skills.

Nice to have:

• Hands-on experience with post-training techniques.
• Experience with multimodal or low-data training.
• Familiarity with large-scale model training.

#LI-KP1