About
JetBrains develops tools that automate routine checks and corrections to help developers work more efficiently. Its AI platform provides capabilities across JetBrains products, using in-house models for writing and coding assistance and integrations with strategic partners.
Responsibilities
•Work with stakeholders to turn business requirements into technical specifications
•Train LLMs from scratch on a large GPU cluster
•Collect and process pre-training and fine-tuning datasets
•Support and improve existing subsystems
Requirements
•Experience designing, deploying, and supporting production ML systems
•Strong theoretical background in NLP and transformer-based approaches
•Proficiency with modern deep learning frameworks such as PyTorch and common NLP libraries
•Experience with distributed training of multi-billion-parameter models
•Attention to detail and strong communication skills
•Ability to plan projects and make decisions independently, consulting others when needed
•Ability to identify customer needs and prioritize tasks accordingly
•Ability to start with simple solutions and add complexity as needed
•Ability to take sole responsibility for an entire subsystem
•Passion for learning and staying up to date with developments in the LLM field
Nice to have
•Experience with LLM inference frameworks such as vLLM, DeepSpeed, and TensorRT
•LLM alignment techniques such as RLHF/RLAIF
•MLOps tools and practices, including CI/CD for ML
•K8s and Kubeflow
•Scientific publications in NLP
