You will be responsible for developing large-scale AI cloud services, including GPU virtual machines, managed Kubernetes platforms, and distributed training pipelines. Additionally, you will work on AI Agent platforms, high-performance networking, and automated operation systems for global AI infrastructure.
Design and operate multi-tenant SDN and high-performance GPU interconnect fabrics for large-scale AI training. Collaborate with compute and control plane teams to automate network provisioning and ensure secure, low-latency infrastructure.
Integrate and operate distributed and parallel file systems in the GPU cloud, managing provisioning, mounting, tenant isolation, quotas, lifecycle, performance, and multi-region data reliability. Own the golden image, GPU driver, CUDA, and container registry pipeline, while collaborating with compute, network, and control-plane teams and supporting monitoring and capacity planning.
Design and build control-plane services for tenant management, quotas, billing and metering, ticketing, permissions, and customer-facing portals and APIs. Develop global and regional platform architecture, self-service onboarding, operational recovery tools, and region-aware scheduling and placement.
You will own the quality gate process, including regression, compatibility, and release validation to ensure fast and stable platform deployments. Responsibilities include automating validation pipelines, managing the compatibility matrix, and defining release acceptance criteria.
Own the end-to-end lifecycle of bare-metal GPU nodes across regions, including provisioning, delivery, operations, break-fix, firmware and DPU management, and decommissioning. Build automated delivery pipelines and reliability practices, lead incident response, and coordinate with storage, image, and network teams to improve fleet availability and capacity.
Own end-to-end reliability of the customer-facing GPU cloud, including Kubernetes operations, tenant lifecycle and isolation, bare-metal provisioning, observability, incident response, and customer communications. Define and meet customer SLAs/SLOs, automate GPU fault remediation and workload recovery, and build self-service operational capabilities for tenants.
You will develop and architect large-scale AI cloud services, including GPU infrastructure, Kubernetes platforms, and distributed training systems. The role involves building enterprise AI Agent platforms and optimizing high-performance networking and model-serving capabilities.
You will own the IP/Ethernet infrastructure connecting the global GPU cloud, including DCI, WAN, and internet edge networks. Additionally, you will build automation to transition network operations from manual tickets to intent-driven policy as code.
The role involves designing and maintaining CI/CD and MLOps pipelines while managing cloud-native infrastructure using Kubernetes and Docker. You will also act as a technical lead for incident management and ensure high availability and security across AI-powered platforms.
You will architect and maintain high-performance storage solutions to support AI model training and inference workloads. This involves managing container storage interfaces, optimizing I/O throughput, and ensuring seamless data delivery for GPU-intensive applications.
You will architect and maintain the high-performance hardware foundation for AI-native cloud computing by integrating GPU device plugins and optimizing network stacks. Additionally, you will manage automated hardware remediation pipelines and tune kernel-level parameters to ensure low-latency, high-bandwidth performance for distributed AI workloads.
You will architect and build a highly available, automated control plane for the NeoCloud platform to manage thousands of Kubernetes clusters. This involves developing custom operators, optimizing etcd performance, and ensuring seamless multi-cluster federation for AI workloads.
Design and implement advanced batch scheduling architectures to optimize AI workload placement and resource utilization. Collaborate with hardware teams to integrate scheduling layers with bare-metal infrastructure and manage complex accelerator requests.
You will contribute to the development of the NeoCloud SRE platform by building monitoring, storage, and automation components under the guidance of senior engineers. This involves writing production code, implementing observability instrumentation, and participating in on-call rotations to maintain system reliability.
Architect and scale a high-cardinality telemetry infrastructure to monitor massive-scale AI hardware and software performance. Develop automated diagnostic tools and alerting pipelines to ensure real-time detection and resolution of system bottlenecks.
The Staff Backend Engineer will co-own the end-to-end design and implementation of a globally distributed, multi-tenant Model-as-a-Service platform. This role involves driving technical strategy, optimizing performance for high-throughput AI inference, and ensuring robust reliability and billing accuracy.
You will deploy and operate high-performance parallel storage systems to support AI training and inference workloads. Additionally, you will instrument storage telemetry and partner with the platform team to develop predictive models for storage-fault detection.
You will design, deploy, and operate the control plane for AI-operated GPU cloud infrastructure, ensuring high availability and automated remediation. You will manage production Kubernetes clusters at scale, implementing topology-aware scheduling and automated failure handling for GPU workloads.
You will lead the end-to-end architecture design of AI Data Center Networks, high-performance Data Center Interconnects, and global backbone networks. Additionally, you will collaborate with engineering teams to implement and optimize hyperscale cluster network standards and configurations.
