Staff Research Engineer (LLM Pre-Training)

JetBrains seeks a Staff Research Engineer to develop and train large language models for coding tasks. The role requires designing technical specifications, managing pre-training datasets, and deploying models on large GPU clusters. Candidates must possess a strong background in NLP, transformer architectures, and distributed training of multi-billion parameter models using PyTorch. Experience with production ML systems is essential, while familiarity with inference frameworks like vLLM and MLOps tools is preferred. This position supports the integration of AI capabilities into JetBrains products, requiring independent decision-making and a focus on efficient, scalable solutions.

Key details

• Role involves training large language models from scratch for coding assistance.
• Candidates need experience with distributed training of multi-billion parameter models.
• Required skills include PyTorch, NLP theory, and production ML system deployment.
• Preferred experience includes vLLM, DeepSpeed, and MLOps practices.
• Positions are available in multiple European cities and remote locations in Germany.