Job Responsibilities:

1. Distributed Training Engineering

• Participate in the implementation of large-scale distributed training solutions.
• Lead the engineering deployment of data parallelism, model parallelism (TP/PP), and ZeRO optimization.
• Continuously tune GPU compute utilization and ensure stability of ultra-large-scale training tasks.

2. Compute Scheduling Optimization

• Deeply involved in the development and optimization of AI task scheduling logic.
• Implement fine-grained resource management, fault self-healing, and efficient Checkpoint mechanisms.
• Solve compute bottlenecks in complex gaming scenarios.

3. End-to-End Model Engineering

• Own the full pipeline from model training to inference deployment.
• Participate in operator performance profiling, model quantization, and high-performance inference pipe...