Deep Learning Infrastructure Engineer at beebom

About the role

Optimize deep learning training at scale. Work on distributed training, GPU cluster management, and performance optimization for massive neural networks.

Responsibilities

• Optimize distributed training across GPU clusters
• Implement mixed-precision training and gradient compression
• Manage and optimize GPU/TPU infrastructure
• Debug performance bottlenecks in training pipelines

Requirements

• BS/MS in Computer Science
• 3+ years experience in distributed systems
• Expert in CUDA, cuDNN, NCCL
• Experience with distributed training frameworks (DeepSpeed, Horovod)
• Strong proficiency in C++ and Python

Benefits

• Top-tier compensation
• Latest hardware access
• Stock options
• Unlimited PTO