AI异构硬件训练优化专家-Seed
Core
Optimizing performance and stability for large-scale LLM training on custom AI accelerator clusters for ByteDance's Seed team.
Role type
Senior IC AI hardware training optimization engineer
Builds
High-throughput, stable training runs for proprietary LLMs supporting 50+ applications (e.g., Doubao, Jimeng)
Domain
AI Infrastructure / High-Performance Computing / Distributed Systems
Deliverable
production ML models
Required skills
C/C++ or Python, Linux, Computer Architecture, Chip Microarchitecture, High-Performance Computing, Distributed Systems, Parallel Computing
Preferred skills
GPU/NPU cluster optimization, NCCL/HCCL/RDMA/RoCE, MoE training, Mixed-precision training, ZeRO, Checkpoint optimization, Fault tolerance, Silent data corruption detection
Technologies
PyTorch, Megatron, DeepSpeed, FSDP, TorchTitan, NCCL, HCCL, RDMA, RoCE
Responsibilities
Optimize data, tensor, pipeline, and expert parallelism strategies; Resolve network jitter, communication bottlenecks, and silent data errors; Collaborate on compiler stacks and compute architectures for cluster stability.
Seniority
Senior, hands-on IC
