About the job
At JetBrains, our unwavering passion for code drives us. Since our inception in 2000, we've dedicated ourselves to minimizing mundane tasks so developers can focus on innovation and delivery.
The JetBrains Research team is at the forefront of exploring machine learning methodologies and agent-based strategies for software development. Our mission is to pioneer the next generation of machine learning capabilities for code functionalities. We invite a talented researcher to enhance and develop training methodologies for models designed to tackle challenges in code generation, code editing, and empowering AI agents. While expertise in every area isn't a prerequisite, comfort in navigating diverse topics and producing publishable outcomes is essential.
Key Responsibilities:
- Train and scale neural networks for analyzing and processing source code.
- Innovate and refine techniques for code generation, code editing, and agent-based workflows.
- Guide and mentor colleagues on machine learning subjects.
Qualifications:
- Proven experience in natural language processing or code modeling.
- Hands-on experience with distributed training and scaling neural networks.
- Familiarity with contemporary training techniques and architectures, such as Mixture of Experts (MoE).
- Published research in leading machine learning conferences.
- Strong programming skills across relevant languages.
Preferred Qualifications:
- Practical experience with post-training techniques.
- Background in multimodal or low-data training methodologies.
- Knowledge of large-scale model training processes.

