Qubrid AI · India
Read everything carefully. The requirements and screening questions are critical. Applications that do not answer them correctly and satisfactorily will be automatically rejected.
Work from home.
Full-time role: this position requires complete commitment. Candidates planning to work two or more jobs simultaneously, or seeking part-time work, should not apply - such applications will be automatically rejected.
Working hours: late nights in India, until approximately 4 AM IST, to overlap with US working hours. Do not apply if this schedule does not work for you.
Compensation: depends on experience and current verifiable compensation (paychecks).
Experience: junior candidates with 2+ years of strong hands-on experience are suitable.
Qubrid AI is building a full-stack AI infrastructure platform that combines GPU cloud, inference APIs, AI orchestration software, and enterprise AI infrastructure. Our platform powers AI workloads across cloud, hybrid, and on-prem environments using state-of-the-art NVIDIA technologies and open-source AI frameworks.
We are looking for a hands-on GPU Kubernetes Cluster Engineer with deep Linux, Kubernetes, cluster automation, and Helm package development expertise to deploy, manage, automate, and optimize GPU clusters used for AI training and large-scale inference workloads.
This role is for someone who has operated real production Kubernetes clusters - not just run local demos or followed tutorials. You will build and run highly available Linux-based GPU clusters, write the Helm charts and automation that deploy them, and own reliability across the stack.
The ideal candidate is comfortable working across:
Production Kubernetes cluster deployment and administration
Linux-based GPU clusters for training and inference
Helm chart and package development for platform and AI workloads
Cluster automation for provisioning, upgrades, and day-2 operations
NVIDIA GPU Operator, drivers, CUDA, and container runtimes
Container networking, ingress, and high-performance networking
Persistent and distributed storage
Monitoring, logging, and alerting
GitOps and Infrastructure as Code
You will work closely with GPU engineers, AI platform teams, and customers to deliver reliable, scalable infrastructure for AI workloads.
Deploy, configure, and maintain production Kubernetes clusters for AI and GPU workloads
Bootstrap and manage multi-node clusters using kubeadm
Configure Kubernetes scheduling, namespaces, resource quotas, node pools, taints, tolerations, and affinity rules
Design and operate multi-tenant Kubernetes environments
Configure high availability and perform cluster upgrades with minimal downtime
Manage container runtimes including containerd and Docker
Configure container networking and CNI plugins such as Calico or Cilium
Manage ingress controllers, load balancing, DNS, and TLS certificates
Design, write, and maintain production-grade Helm charts for platform services and AI workloads
Build reusable, parameterized charts with clean values structures, templates, helpers, and dependencies
Package GPU-enabled workloads (inference servers, model services, operators) as versioned Helm releases
Manage chart versioning, chart repositories, and release lifecycles (install, upgrade, rollback)
Implement chart testing, linting, and validation before release
Deploy and manage workloads using Helm, Operators, and GitOps methodologies
Automate cluster provisioning, configuration, and scaling end to end
Automate Kubernetes upgrades, node onboarding, and day-2 operations
Build automation using Ansible, Bash, and Python
Implement GitOps workflows for declarative cluster and application management
Apply Infrastructure as Code principles to make cluster builds repeatable and auditable
Reduce manual operational work through scripts, pipelines, and self-healing patterns
Build and operate Linux-based GPU clusters supporting training and inference environments
Install and manage NVIDIA GPU Operator, drivers, CUDA, and GPU-enabled container runtimes
Configure GPU scheduling, sharing, and partitioning for AI workloads
Collaborate with GPU engineers and AI platform teams to optimize infrastructure for inference and training
Manage persistent storage solutions and distributed file systems
Troubleshoot Kubernetes, Linux, networking, and storage issues across distributed systems
Implement cluster monitoring, logging, and alerting
Diagnose and resolve performance bottlenecks in GPU and cluster environments
Participate in customer deployments across cloud, hybrid, and on-prem environments
Provide operational support for production clusters
Write technical documentation, runbooks, and deployment guides
Take ownership of infrastructure from initial build through production operation and maintenance
4+ years of Linux systems administration experience
3+ years of hands-on Kubernetes administration experience
Strong expertise with:
Kubernetes
kubeadm
containerd
Helm (including writing and maintaining your own charts)
Docker
Linux (Ubuntu, Rocky Linux, RHEL)
Proven experience developing Helm charts and packaging applications for Kubernetes
Experience automating cluster provisioning and operations
Experience deploying and managing multi-node clusters
Experience with NVIDIA GPU Operator and GPU-enabled Kubernetes environments
Solid understanding of:
Networking fundamentals
DNS
Load balancing
Ingress controllers
TLS certificates
High availability
Experience with container networking and CNI plugins such as Calico or Cilium
Strong scripting skills in Bash and Python
Familiarity with Git and Infrastructure as Code principles
Ability to troubleshoot distributed systems and performance issues
Experience with GPU clusters and AI infrastructure
Knowledge of NVIDIA drivers, CUDA, MIG, and DCGM
Experience with KubeVirt and GPU virtualization
Familiarity with Ceph, Longhorn, OpenEBS, NFS, or other persistent storage solutions
Experience with monitoring and observability tools including:
Prometheus
Grafana
Loki
Alertmanager
Experience with ArgoCD, FluxCD, and GitOps workflows
Experience with Kubernetes Operators or custom controllers
Understanding of RDMA, RoCEv2, InfiniBand, and high-speed networking
Familiarity with Slurm and HPC environments
Experience with Rancher, OpenShift, or VMware Tanzu
Exposure to AI inference frameworks such as Triton, vLLM, SGLang, and TensorRT-LLM
CKA (Certified Kubernetes Administrator), CKS, or RHCE certifications
Experience supporting H100, H200, B200, A100, or HGX-based clusters
Familiarity with SONiC networking and NVIDIA Spectrum switches
Experience operating clusters with hundreds or thousands of GPUs
Infrastructure automation using Terraform or Ansible
Experience with hybrid cloud environments
Large-scale GPU clusters for AI training and inference
Kubernetes-based AI platforms and multi-tenant environments
Helm charts and packaging for AI platform services and GPU workloads
Cluster automation and GitOps-driven operations
NVIDIA GPU Operator and containerized AI workloads
Distributed storage and high-performance networking
Hybrid cloud and on-prem AI deployments
Enterprise AI factories and next-generation GPU cloud infrastructure
Someone who has run production Kubernetes clusters, not just deployed to them
Someone who writes Helm charts and automation, not just installs existing ones
A strong troubleshooter who enjoys debugging complex distributed systems
An automation-first mindset - if it's done twice, it gets scripted
An ownership mentality with strong execution skills
Someone comfortable working across Linux, Kubernetes, networking, storage, and GPUs without heavy supervision
Someone who understands that reliable AI infrastructure requires more than spinning up a cluster - it requires thoughtful automation, packaging, observability, and operational discipline
At Qubrid AI, you'll help build the infrastructure behind the next generation of AI. You'll work with cutting-edge GPU clusters, Kubernetes, and NVIDIA technologies to power AI applications at scale for enterprises and developers worldwide.
If you enjoy Linux, Kubernetes, automation, and solving complex infrastructure challenges, we'd love to hear from you.
