Staff Research Engineer LLM Pre-Training

Jetbrains

Staff Research Engineer LLM Pre-Training Overview

Company Name Jetbrains
Job Role Staff Research Engineer LLM Pre-Training
Qualifications Not Specified
Category IT Jobs
Job Type Full Time
Location London

JetBrains is seeking a highly skilled Staff Research Engineer specializing in Large Language Model (LLM) pre-training to join their innovative team. The company has been dedicated to creating powerful developer tools since 2000 and is now expanding into AI capabilities integrated across its product suite. This role involves working on developing foundational models from scratch, deploying them into production, and supporting their ongoing improvement.

Key Responsibilities

  • Work closely with stakeholders to convert business requirements into detailed technical plans for LLM projects.
  • Train large-scale language models from the ground up using extensive GPU clusters, ensuring optimal performance and accuracy.
  • Collect, preprocess, and manage datasets for pre-training and fine-tuning models to ensure high-quality training data.
  • Support and enhance existing subsystems related to LLM development, deployment, and maintenance.
  • Deploy trained models into production environments, making them accessible to users worldwide.
  • Keep up-to-date with the latest developments in NLP and transformer architectures, applying new techniques to improve models.
  • Contribute to the development of AI features integrated into JetBrains products, supporting the company’s strategic AI platform.

Candidate Requirements

  • Experience in designing, deploying, and supporting machine learning systems in a production environment.
  • Strong theoretical background in NLP and transformer-based approaches.
  • Proficiency with deep learning frameworks such as PyTorch and relevant NLP libraries.
  • Experience with distributed training of models with billions of parameters.
  • Excellent attention to detail and communication skills for effective teamwork.
  • Ability to independently plan projects, make decisions, and prioritize tasks based on customer needs.

Preferred Skills and Experience

  • Experience with inference frameworks like vLLM, DeepSpeed, or TensorRT.
  • Knowledge of LLM alignment techniques such as RLHF or RLAIF.
  • Familiarity with MLOps practices, including CI/CD pipelines for ML workflows.
  • Experience with Kubernetes (K8s) and Kubeflow.
  • Publication record in NLP research is a plus.

Development Environment and Tools

  • Access to a large GPU cluster of NVIDIA hardware for training models.
  • Version control with Git.
  • ML stack including Python, PyTorch, and HuggingFace.
  • Experiment tracking with Kubeflow and Weights & Biases.
  • Continuous integration managed via TeamCity.

Workplace and Culture

JetBrains fosters an inclusive and open work environment, welcoming individuals from all backgrounds, identities, and locations. The company offers flexible working arrangements, including remote options across various European locations, and is open to discussing relocation for suitable candidates.

Application Process

Interested candidates can apply by submitting their personal details, CV, cover letter, and any relevant links such as LinkedIn or GitHub. The application process is managed through an online platform, and the company emphasizes equal opportunity employment.

Note: The company explicitly states that they offer visa sponsorship for this role.


Degree Requirement: Not Specified

Visa Sponsorship Promising

To apply for this job please visit job-boards.eu.greenhouse.io.