Research Engineer (Agentic Models)

About the role

JetBrains is building AI‑powered coding agents that can understand large codebases, plan changes, call tools and iterate with developers inside its IDEs. As a Research Engineer on the Agentic Models team you will design, implement and maintain the training and evaluation pipelines that power these multi‑step agents.

Key responsibilities

• Design and operate supervised‑fine‑tuning (SFT) and reinforcement‑learning (RL) post‑training pipelines for coding agents.
• Adapt large language models (LLMs) for agent workflows, including planning, tool usage and multi‑step interactions.
• Build simulation and evaluation environments to measure agent performance on realistic developer tasks.
• Define metrics, analyse logs and close the loop from evaluation back to training, data and reward design.
• Improve model architectures, training recipes and datasets based on experimental results.
• Work with large‑scale GPU clusters and MapReduce‑style data pipelines for pre‑training and fine‑tuning.
• Collaborate with research, product and infrastructure teams to turn product visions into shipped features.

Required profile

• 3+ years of Python development in modern ML codebases.
• Hands‑on experience training LLMs (pre‑training, fine‑tuning or post‑training) in research or production.
• Deep knowledge of PyTorch and LLM training stacks such as Megatron, NeMo or similar.
• Strong understanding of LLM fundamentals, distributed training, mixed‑precision and debugging unstable runs.
• Product‑aware mindset with ability to translate developer needs into modeling work.

Required skills

• Python
• PyTorch
• Megatron
• NeMo
• verL (or similar LLM training frameworks)
• Kubeflow, Dagster, Airflow, ZenML
• Kubernetes, SLURM
• MapReduce‑style data processing
• GPU clusters and distributed training