New opportunity

AI Infrastructure Engineer

vCluster Labs · Remote; Singapore

About this role

As an AI Infrastructure Engineer at vCluster Labs, you will work directly with customers at the earliest and most critical stage of their journey, from bare metal GPU nodes through to production-ready deployment. This is a technical role operating pre-sale as part of a proof of value engagement. You will help GPU AI Clouds and enterprises building AI Factories to offer Kubernetes as a managed service on bare metal GPU infrastructure. Responsibilities include leading end-to-end technical deployments, configuring and troubleshooting bare metal GPU node infrastructure, validating Kubernetes and vCluster environments, and facilitating knowledge transfer to customer teams. You will also document reusable playbooks, collaborate with Engineering and Product to surface infrastructure challenges, and partner with Sales during the pre-sales process. Required qualifications include 5+ years of experience deploying and operating Kubernetes in production, ideally on bare metal, along with practical knowledge of NVIDIA GPU Operators, CUDA tooling, and systems-level configuration for GPU nodes. You must possess deep understanding of networking fundamentals (CNI, overlay networks, load balancing) and storage expertise (CSI drivers, distributed systems like Ceph, Rook, Weka, or Longhorn). Operational agility and a modern tech mindset are essential. Bonus points for automation skills (Bash, Python, Go), CKA certification, experience writing Kubernetes Operators, and familiarity with AI/ML tooling such as inference serving, GPU scheduling, and LLM deployment. The role is remote-first and offers a competitive salary, equity, and comprehensive benefits.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

KubernetesBare metal configurationNVIDIA GPU OperatorsCUDACNI pluginsOverlay networksLoad balancingPersistent volume configurationCSI driversCephRookWekaLonghornBashPythonGoCKAKubernetes OperatorsInference servingGPU schedulingLLM deploymentRDMAInfiniBand

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

End-to-End Bare Metal GPU Deployment
The core of the role is driving technical deployments from raw hardware to a functional Kubernetes environment. You must demonstrate proficiency in configuring bare metal nodes, including NVIDIA GPU Operators, CUDA tooling, and complex networking (RDMA/InfiniBand) to ensure high-performance AI workloads.
Pre-Sales Technical Validation
As this role operates during the proof-of-value phase, you must bridge the gap between sales and engineering. You are expected to validate the vCluster platform's viability for customers, requiring both deep technical troubleshooting skills and the ability to communicate value effectively to stakeholders.
Scalable Knowledge Documentation
Because you are building the foundation for future deployments, you must prioritize creating reusable playbooks and architectural documentation. This ensures that your field learnings directly contribute to the company's ability to scale its customer onboarding process.

Likely challenges to think through

High-Stakes Ambiguity in Proof-of-Value
Operating in a pre-sales environment means you will encounter unique, non-standard customer infrastructure configurations. Prepare by practicing how to diagnose and resolve issues in unfamiliar environments under tight deadlines, as you will often be writing deployment playbooks in real-time.
Balancing Field Feedback with Product Roadmap
You are the primary conduit between customer infrastructure pain points and the internal engineering team. The challenge lies in distilling complex, site-specific technical hurdles into actionable product feedback without losing sight of the immediate customer delivery goals.

Practice your answers

Technical practice

Bare Metal GPU Node Configuration

What to coverExplain the process of preparing a bare metal node for Kubernetes, specifically focusing on GPU driver installation, CUDA toolkit setup, and ensuring proper hardware visibility for the NVIDIA GPU Operator. Suggestion: Prepare to discuss how you handle firmware updates and kernel-level configurations that might impact GPU performance.

Kubernetes Networking and CNI

What to coverDescribe how you would troubleshoot connectivity issues in a complex, layered Kubernetes environment involving overlay networks and load balancers. Suggestion: Focus on your experience with CNI plugin selection and how you diagnose traffic flow between pods and external services.

Distributed Storage for AI Workloads

What to coverCompare the trade-offs between different distributed storage backends like Ceph, Rook, or Weka when supporting high-throughput AI/ML training workloads. Suggestion: Discuss how you configure CSI drivers to ensure persistent volume performance meets the demands of GPU-intensive applications.

NVIDIA GPU Operator and Scheduling

What to coverHow do you manage GPU scheduling in Kubernetes to ensure efficient resource allocation for multiple tenants? Suggestion: Explain your approach to using the NVIDIA GPU Operator and how you handle pod-to-GPU mapping in a multi-tenant environment.

RDMA and InfiniBand Integration

What to coverWhat are the primary challenges when configuring RDMA and InfiniBand for high-performance computing clusters on Kubernetes? Suggestion: Highlight your experience in ensuring low-latency communication between nodes and how you validate these configurations.

Kubernetes Operator Development

What to coverDescribe a scenario where you would choose to build a custom Kubernetes Operator rather than using existing tooling. Suggestion: Focus on how you manage the lifecycle of complex stateful applications and the benefits of automating these processes.

Inference Serving and LLM Deployment

What to coverWhat are the key infrastructure considerations for deploying LLMs at scale, particularly regarding inference serving and resource isolation? Suggestion: Discuss how you balance latency, throughput, and GPU memory usage when designing these deployment architectures.

Automation and Playbook Creation

What to coverHow do you approach documenting infrastructure deployments to ensure they are reusable and scalable for future customers? Suggestion: Provide examples of how you use scripts (Bash, Python, or Go) to automate repetitive tasks and reduce human error.

Behavioural practice

Handling Ambiguity in Pre-Sales

What to coverDescribe a time you had to deliver a technical proof of value in an ambiguous or fast-moving environment. Suggestion: Use the STAR method to explain how you identified the customer's core requirements, managed expectations, and successfully reached a production-ready state.

Customer Knowledge Transfer

What to coverGive an example of how you enabled a customer team to become self-sufficient with a complex platform you deployed. Suggestion: Focus on your communication style, the documentation you created, and how you measured their success in operating the platform independently.

Collaborating with Product and Engineering

What to coverTell me about a time you identified a recurring infrastructure challenge in the field and successfully influenced the product roadmap. Suggestion: Explain how you gathered data, communicated the impact to internal stakeholders, and helped drive a solution.

Ownership and Delivering Value

What to coverDescribe a situation where you went beyond the initial scope of a project to ensure the customer achieved their desired outcome. Suggestion: Highlight your 'Own the Outcome' mindset and how you prioritized actions that provided long-term value for the business and the client.

Sharpen your knowledge

Topics to prepare

Bare Metal Kubernetes Architecture
Focus on the complexities of running Kubernetes directly on hardware without hypervisors. Exercise: Design a deployment strategy for a multi-node bare metal cluster that accounts for PXE booting, node lifecycle management, and hardware-level isolation. This is critical for the role's focus on enabling AI Factories to offer managed services on bare metal.
GPU-Accelerated Networking and Storage
Study the integration of RDMA/InfiniBand with Kubernetes and high-performance distributed storage systems like Ceph or Weka. Exercise: Map out how CNI plugins and CSI drivers interact to provide low-latency data paths for GPU workloads. This is essential for troubleshooting the infrastructure optimization tasks mentioned in the job description.

Skills to revise

NVIDIA GPU Operator Configuration
Gain hands-on experience deploying and tuning the NVIDIA GPU Operator within a Kubernetes cluster. Exercise: Create a lab environment to automate the installation of GPU drivers, monitoring tools, and device plugins, then validate that a pod can successfully request and utilize GPU resources. This is a core requirement for GPU fluency.
Kubernetes Operator Development
Develop a custom controller or operator to automate infrastructure tasks. Exercise: Write a simple Go-based operator that watches for a custom resource and triggers a configuration change on a node. This aligns with the bonus qualification for Kubernetes depth and the need to build reusable playbooks.
Infrastructure Automation Scripting
Practice automating repetitive deployment tasks using Bash or Python. Exercise: Write a script that automates the validation of a node's readiness, including checking for CUDA availability, network connectivity, and storage mount points. This is directly relevant to the role's requirement for scaling through documentation and playbooks.

Your preparation checklist

Document a complex bare metal Kubernetes deployment
Prepare a technical walkthrough of a production-grade Kubernetes cluster you deployed on bare metal. Highlight how you handled CNI configuration, storage backends like Ceph or Rook, and node-level troubleshooting to demonstrate your infrastructure expertise.
Prepare a GPU infrastructure case study
Create a brief summary of your experience with NVIDIA GPU Operators and CUDA tooling. Focus on a specific scenario where you optimized GPU node performance or resolved a scheduling bottleneck, as this directly maps to the role's core responsibilities.
Draft a reusable technical playbook
Write a sample 'getting started' or 'troubleshooting' guide for a complex infrastructure component. This demonstrates your ability to scale knowledge for customers, a key requirement for the role's focus on documentation and customer self-sufficiency.
Formulate questions on pre-sales technical engagement
Prepare thoughtful questions about how the team balances proof-of-value engagements with long-term product feedback loops. Ask how they measure success during the pre-sale phase to show you understand the commercial impact of your technical work.
Review your automation and scripting portfolio
Organize examples of Bash, Python, or Go scripts used to automate infrastructure tasks. Be ready to explain how these scripts improved operational agility or reduced manual configuration time in previous high-complexity environments.

Availability last checked: 18 Oct 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in Singapore. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…