X
【27届校招】大模型训练推理框架工程师
Xpeng 2
深圳、北京
≈ 235,8 тыс.–471,7 тыс. ₽ · 2,9 тыс.–5,9 тыс. $ · 2,6 тыс.–5,1 тыс. €
20 000 – 40 000 CNY
🏢 Офис
Middle
Полная занятость
Продуктовая
Tech
Описание вакансии
Обязанности
•Build distributed reinforcement learning training environment
•Build evaluation platform for LLM VLM and multimodal models
•Design training communication scheduling and fault tolerance
•Develop MLOps toolchain for model versioning and experiment tracking
•Develop distributed training infrastructure
•Integrate evaluation reporting with CI CD
•Optimize large model inference deployment
Технологии: Airflow, C++, CI/CD, CUDA, DeepSpeed, FSDP, Go, High Performance, High-Performance Computing, Kubernetes, MLOps, Megatron-LM, NCCL, Performance Computing, Python, Quantization, RDMA, Ray, Reinforcement Learning, TensorRT, VLLM