AI Infra工程师
Tasks
- Build training profiling and diagnostics
- Design data loading pipelines
- Develop standardized training frameworks
- Identify and resolve training performance bottlenecks
- Implement distributed training strategies
- Improve CPU GPU data pipeline
- Integrate multi cloud ML platform capabilities
- Optimize PyTorch DataLoader and IO
- Platformize training infrastructure for reproducibility and stability
Perks/Benefits
- N/A
Skills/Tech-stack
AMP | BF16 | Batch Size | CPU Decoding | CUDA | Ceph | Checkpoint | DDP | DataLoader | DataSet | DeepSpeed | Docker | FSDP | File System | GPU Training | GPU Training Optimization | GPU Utilization | Gradient Accumulation | HDF5 | Iostat | JPEG | JuiceFS | Kubernetes | Linux | MFU | Megatron-LM | NFS | NVIDIA GPU | NVIDIA Nsight | NVIDIA Nsight Compute | NVIDIA SMI | Nsight Compute | Nsight Systems | NumPy | Nvidia NSight Systems | OSS | POSIX File System | Parquet | Perf | Profiling | PyTorch | PyTorch Profiler | Python | Ray Train | S3 | Strace | TCP-Dump | TCPDump | Torch Distributed | Training Optimization | WebDataset
Education
Bachelor of Engineering | Bachelor of Science | Master of Science
Related jobs
-
ARM | Android | Communication Protocols | Core Architecture | DebuggingSenior-level Full TimeShanghai, China5h ago
-
Senior-level Full TimeShanghai, China5h ago
-
Automation & Robotics Engineer CNY 360K-600KActuator integration | Analog Signal | Artificial Intelligence | C++ | CADSenior-level Full TimeHangzhou, China16h ago
-
Entry-level Internship上海19h ago
-
Senior-level Full Time上海、北京20h ago
-
Entry-level Internship上海20h ago
-
Mid-level Full Time上海20h ago
-
Mid-level Full Time上海、深圳 R20h ago
-
Mid-level Full Time上海20h ago
-
Senior-level Full Time上海20h ago
-
Mid-level Full Time北京、上海20h ago
-
Entry-level Full Time上海20h ago
-
Entry-level Full Time北京、上海20h ago
-
Mid-level Full Time上海20h ago
-
Mid-level Full Time上海、北京20h ago
-
Mid-level Full Time上海20h ago
-
Senior-level Full Time上海20h ago
-
3D视觉算法工程师 CNY 180K-300K3D Pose Estimation | 3D Reconstruction | 6D Pose | 6D Pose Estimation | Audio ProcessingMid-level Full Time深圳20h ago
-
Senior-level Full Time上海 R20h ago
-
Mid-level Full Time上海 R20h ago
-
Senior-level Full Time上海、北京20h ago
-
Senior-level Full Time上海 R20h ago
-
loco-manipulation(高动态物体交互方向) CNY 180K-360KBody Dynamics | C++ | Contact dynamics | Curriculum learning | Domain RandomizationMid-level Full Time上海20h ago
-
Data Infra高级研发工程师(AI与具身智能数据平台方向) CNY 144K-240KAsynchronous processing | Automated testing | CI/CD | CRD | Cloud NativeSenior-level Full Time上海 R20h ago
-
Mid-level Full Time北京、上海20h ago