Research Engineer - (Seed Model - Infra - Pretraining Infrastructure)

What you'll do

• Design and optimize distributed training strategies, including parallelism and computation and communication optimization on large GPU clusters.
• Analyze bottlenecks in large-scale training systems and develop data-driven optimizations.
• Translate research ideas into scalable production infrastructure.

What they ask for

• Bachelor's degree or above in Computer Science, Software Engineering, or a related field.
• 2+ years of experience in large-scale model training, distributed training systems, or GPU kernel optimization.
• Solid understanding of large-scale LLM and multimodal model training, with experience in GPU-cluster parallelism strategies or GPU kernel performance optimization.
• Strong grasp of algorithm–system co-design and cross-layer optimization, plus strong programming skills and hands-on experience implementing production-grade ML systems or infrastructure.

Nice to have

• Hands-on experience pretraining LLMs, multimodal, or video generation models at scale; using mainstream distributed training frameworks; or developing custom kernels with CUDA or Triton.
• Publications at top-tier machine learning or systems conferences, or notable open-source contributions to widely used training frameworks.

Tools

Megatron-LM, DeepSpeed, PyTorch FSDP, NCCL, CUDA, Triton

AI focus

Infrastructure for efficient pretraining of foundation models, multimodal LLMs, and video generation models.