推理基础设施架构师（远程） at Telnyx

部门：Product

语言

工作方式: 远程
地点: 中国
职位类型: 全日制
级别: staff+
发表: 2026年9月16日

说明

Telnyx 正在为其中国创始团队招聘一名高级或资深级推理基础设施架构师，负责运营并扩展裸金属 B300 GPU 集群，并构建无服务器和专用 LLM 推理平台。该职位涵盖 vLLM 和 SGLang 服务、Kubernetes GPU 调度、模型路由、KV 缓存、权重分发、自动扩缩容、可观测性和成本优化，并需为开源基础设施贡献上游代码。候选人应具备生产环境 LLM 服务经验、Kubernetes 和 GPU 集群运维专长，以及出色的性能工程技能；该职位可在中国大陆远程工作，并可为搬迁提供签证赞助。

所需技能

vLLM
SGLang
Kubernetes
GPU Operator
Kueue
Python
Go
OpenTelemetry

附加信息

Required Previous Experiences: Production LLM serving under meaningful traffic and latency constraints
End-to-end Kubernetes GPU fleet operations
Deep operational command of vLLM or SGLang
System-level performance engineering
Preferred Qualifications: Experience operating inference platforms at large-scale technology companies
Experience with listed LLM infrastructure technologies
CNCF or OpenInfra community involvement
Fine-tuning or reinforcement-learning infrastructure
Real-time voice latency work
Multi-region deployments and data-residency requirements
Translation Company Name: Telnyx

福利和津贴

签证担保

关于 Telnyx

Telnyx

Cloud communicationsAustin, TX, US465

Infrastructure for real-time AI. Edge compute, voice AI agents, and global communications, all on one platform. We combine global telephony, dedicated AI infrastructure, and full customizability under one roof so you can design and deploy AI-powered agents that feel like part of your team. From low-latency voice streaming to scalable, multi-language support, Telnyx gives you everything you need to build real-time, intelligent voice experiences. Whether you're enhancing customer support, automating outbound calls, or embedding voice AI into your product, Telnyx makes it easy to launch and scale with confidence. And you never have to worry about piecing together multiple providers or sacrificing performance. Build agile. Launch faster. Scale globally. All on one platform built for real-time engagement.