机器学习平台研发工程师（深圳/北京/上海）
Core
Design and develop the backend architecture for a Large Model MaaS platform, covering model service deployment, orchestration, and invocation across the full link.
Role type
Senior IC backend engineer (MaaS platform)
Builds
A reusable, scalable, and observable MaaS platform for internal high-volume business model services.
Domain
Cloud-native infrastructure + Large Language Model (LLM) serving
Deliverable
production ML models
Required skills
Java, Golang, Python, Cloud-native technologies (K8s, Docker, ServiceMesh), High-concurrency distributed systems, Microservices, Message queues, Caching, Distributed storage, AI platform development, Inference acceleration frameworks (PyTorch, TensorFlow, VLLM, SGLang)
Preferred skills
Template-based one-click deployment, Heterogeneous card support, Multi-inference framework integration
Technologies
K8s, Docker, ServiceMesh, PyTorch, TensorFlow, VLLM, SGLang
Responsibilities
Design and develop backend architecture for MaaS platform covering deployment, orchestration, and invocation; Optimize system architecture for high availability and performance; Implement template-based one-click deployment for multi-scenario and multi-card types; Enhance platform capabilities and usability for heterogeneous cards and deployment forms.
Seniority
Senior, hands-on IC