Principal Software Engineer - AI Infrastructure
Core
Own the architecture and technical direction of major components of the training platform for Large Language Models (LLMs), Small Language Models (SLMs), multimodal, code-specific, and open-weight models.
Role type
Principal Software Engineer (AI Infrastructure)
Builds
Scalable, reliable, and cost-efficient training infrastructure for proprietary and open-weight models.
Domain
Artificial Intelligence / Machine Learning Infrastructure
Deliverable
production ML models
Required skills
Distributed training frameworks (PyTorch, DeepSpeed, Megatron-LM), GPU cluster scheduling, Reinforcement Learning pipelines, Open-weight model training, Incident response, Technical leadership, Mentorship
Preferred skills
Open-source contributions, AI-assisted engineering tools
Technologies
PyTorch, DeepSpeed, Megatron-LM, Hugging Face, Ray Train, vLLM, SGLang, verl, slime, MILES
Responsibilities
Drive ambiguous, multi-quarter initiatives from proposal to production; Lead debugging and resolution of complex issues in distributed training jobs; Mentor engineers and raise the technical bar through design reviews; Partner with customers to define requirements and prioritize engineering work.
Seniority
Principal, strategy & mentorship
