JetBrains

JetBrains

JetBrains provides cutting-edge development tools like IntelliJ IDEA and Kotlin, automating tasks to boost productivity and foster innovation.

Internet Software & Services
1K-5K
Founded 2000

Description

  • Design, implement, and maintain SFT and RL post-training pipelines for multi-step coding agents.
  • Train and adapt large language models for planning, tool use, and multi-step interactions in JetBrains IDEs.
  • Build simulation and evaluation environments where coding agents can perform and be measured on realistic developer tasks.
  • Design evaluation frameworks and metrics for agent behavior, and use traces and logs to improve training, data, and reward design.
  • Analyze training and evaluation results to improve model architectures, training recipes, and datasets.
  • Work with distributed GPU clusters and MapReduce-style infrastructure for training and data processing.
  • Collaborate with research, product, and infrastructure teams to turn product goals into models, experiments, and shipped features.

Requirements

  • Extensive hands-on experience training LLMs in pre-training, fine-tuning, or post-training settings.
  • Deep expertise in PyTorch and specialized LLM training stacks such as Megatron, NeMo, or verl.
  • Strong understanding of LLM fundamentals, including architectures, tokenization, data pipelines, batching, mixed precision, distributed training, and debugging unstable runs.
  • Ability to own projects end to end from problem definition through design, experimentation, implementation, and iteration.
  • Product-aware mindset with the ability to translate developer needs and failure modes into modeling and evaluation work.
  • At least 3 years of Python experience writing clean, maintainable code in modern ML codebases.
  • Experience with ML orchestrators or workflow tools such as Kubeflow, Dagster, Airflow, or ZenML, or schedulers like Kubernetes or SLURM (preferred).
  • Experience with large-scale data and training pipelines, including MapReduce-style clusters, multi-node GPU training, or workloads around 1M+ CPU/GPU hours (preferred).
  • Experience designing and maintaining evaluation pipelines for LLMs or agents, including metrics, dashboards, experiment tracking, and automated regression checks (preferred).
  • Experience with AI agent development, including tool-using agents, planners, or multi-step coding workflows (preferred).

Interested in this position?

Apply directly on the company website

Apply Now

Similar Roles

Security Engineer - Node.js Proactive Defense (worldwide remote, work anywhere)

CloudLinux 51-250 IT Services

CloudLinux is hiring an engineer-researcher to build a new runtime protection product for Node.js web apps running on managed hosting, from design through production rollout.

JavaScript Node.js PHP WAF WordPress
18 hours, 42 minutes ago

Senior Numerical Optimization Engineer

Anduril Industries 1K-5K Aerospace & Defense

Anduril Industries is hiring a Senior Numerical Optimization Engineer to build scheduling and optimization systems for its Maritime Digital Production team, supporting shipbuilding operations with rapid production replanning.

C++ CI/CD ERP IoT Java Microservices NumPy Pandas Python Rust SciPy
1 day, 18 hours ago

PhD Studentship: Causal Reinforcement Learning

Phaidra 51-250 Internet Software & Services

Phaidra is offering a fully funded 4-year PhD studentship with the University of Cambridge to research causal inference and reinforcement learning for robust AI control systems in industrial and data centre environments.

Machine Learning NumPy Python PyTorch Reinforcement Learning Scikit-learn SciPy
2 days, 18 hours ago

Research Contractor

FUTURESEARCH Internet Software & Services

FutureSearch is hiring a research contractor to help develop tasks for training and benchmarking frontier AI models in a highly research-driven environment.

3 days, 18 hours ago

You're on a roll! Sign up now to keep applying.

Sign Up

Already have an account? Log in

Used by 14,729+ remote workers