机器学习平台SRE工程师（深圳/北京）
Core
Define SLA/SLO, build monitoring/alerting, and ensure stability of the machine learning platform's training and inference chains.
Role type
Senior SRE for Machine Learning Platform
Builds
ML training and inference infrastructure, resource scheduling systems, and automated operational tools
Domain
AI/ML Infrastructure, Distributed Systems, Cloud Native
Deliverable
production ML models | infrastructure
Required skills
Kubernetes, Python/Go/Java/Shell, Distributed Systems, High Availability, Performance Tuning, Database Optimization (MySQL/PostgreSQL), GPU Cluster Management, NCCL/RDMA/InfiniBand, LLM Training/Inference Optimization, Agent/Harness Architecture
Preferred skills
LLM/Multimodal Model Stability, PyTorch/DeepSpeed/Megatron, vLLM/SGLang, MCP (Model Context Protocol)
Responsibilities
Define SLA/SLO and manage fault response for the ML platform; Architect disaster recovery and optimize system performance; Automate operations using AI for resource management and fault recovery.
Seniority
Senior, hands-on IC
