New opportunity

Technology Platform Engineer

Accenture · Gurugram, India

About this role

The Technology Platform Engineer will design and implement high-performance computing (HPC) and AI infrastructure solutions. This role focuses on creating scalable, high-availability environments for AI and HPC workloads, including GPU-as-a-Service and Metal-as-a-Service offerings. Responsibilities include deploying and managing XPU-based clusters, optimizing performance across hybrid cloud environments, and integrating AI platforms with existing IT and security frameworks. Candidates must have at least 7.5 years of experience and a 15-year full-time educational background. Key responsibilities involve designing supercomputing environments, managing large-scale GPU clusters, and providing technical support for distributed training and simulations. Required technical expertise includes Linux architecture, advanced CUDA/GPU management, HPC cluster design (Slurm/PBS Pro), and infrastructure-as-code tools like Terraform and Ansible. Preferred qualifications include experience managing 1,000+ GPU clusters, knowledge of AI frameworks like PyTorch and TensorFlow, and familiarity with Agentic AI for workflow automation. The role requires strong problem-solving skills to deliver robust, scalable infrastructure in fast-paced environments.

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

Skills for this role

Linux ArchitectureHPC infrastructureAI infrastructureGPU cluster managementSlurmKubernetesDockerCUDANVIDIA H100NVIDIA A100InfiniBandRoCEv2MLOpsKubeflowAnsibleTerraformPythonAzure MLMachine LearningAgentic AI

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

Architecting Large-Scale XPU/GPU Clusters
The role requires designing and managing high-performance computing environments, specifically focusing on clusters with 1,000+ GPUs. You must demonstrate expertise in NVIDIA DGX SuperPOD architectures, InfiniBand/RoCEv2 networking, and the ability to scale infrastructure for distributed AI training and simulations.
Infrastructure-as-Code and Automation
You are expected to deploy and manage production environments using automation pipelines. This involves using Terraform and Ansible to maintain configuration baselines and automate the provisioning of Metal-as-a-Service (MaaS) and GPU-as-a-Service (GPUaaS) offerings.
Performance Optimization and Observability
A core priority is ensuring high availability and low-latency performance. You must be capable of tuning HPC/AI workloads, managing cluster schedulers like Slurm, and implementing monitoring frameworks to ensure workload resiliency across hybrid cloud and on-premises environments.

Likely challenges to think through

Managing Complex Hybrid-Cloud Interoperability
The challenge lies in integrating on-premises supercomputing environments with public cloud providers (Azure/AWS) while maintaining consistent security and performance. Prepare by reviewing strategies for hybrid HPC bursting and managing data gravity when moving large-scale AI datasets between environments.
Balancing Cost Efficiency with High-Performance Demands
You will be tasked with optimizing energy and cost efficiency (FinOps) for expensive GPU resources. Prepare by discussing your experience with Reserved Instances, Spot VM strategies, and how you balance the need for maximum compute throughput against strict budgetary constraints.
Technical Support for Diverse AI/HPC User Needs
Providing guidance to users running varied frameworks like PyTorch, TensorFlow, or JAX requires deep knowledge of the full stack. Prepare for scenarios where you must troubleshoot bottlenecks that span across the application layer, container orchestration (Kubernetes/Singularity), and the underlying hardware fabric.

Practice your answers

Technical practice

Designing Large-Scale GPU Clusters

What to coverExplain your approach to architecting a 1,000+ GPU cluster. Focus on considerations for high-performance fabrics like InfiniBand or RoCEv2, multi-rail network designs, and storage solutions like Lustre or GPFS to prevent I/O bottlenecks during distributed training.

HPC Scheduler Optimization

What to coverCompare Slurm and PBS Pro in the context of managing heterogeneous XPU clusters. Discuss how you configure these schedulers to handle job preemption, resource allocation, and fair-share policies for multi-tenant AI and HPC workloads.

Infrastructure-as-Code for HPC

What to coverDescribe how you utilize Terraform and Ansible to automate the deployment of HPC environments. Address how you manage configuration drift and ensure reproducibility across hybrid cloud and on-premises infrastructure.

CUDA and GPU Performance Tuning

What to coverDetail your methodology for troubleshooting performance issues in NVIDIA H100/A100 environments. Discuss the use of profiling tools to identify kernel bottlenecks, memory bandwidth limitations, or inefficient data movement between CPU and GPU.

Containerization in HPC

What to coverDiscuss the trade-offs between using Singularity and Kubernetes for containerizing HPC applications. Explain how you manage container images and runtime environments to ensure high performance and security in a shared cluster.

Hybrid Cloud Bursting Strategies

What to coverExplain the technical challenges of bursting on-premises HPC workloads to public cloud providers like Azure or AWS. Focus on data synchronization, latency management, and cost-optimization strategies such as using Spot instances or Reserved Instances.

AI Platform Integration

What to coverHow do you integrate MLOps tools like Kubeflow or MLflow with existing HPC infrastructure? Describe the process of enabling researchers to run distributed training jobs seamlessly while maintaining security and resource governance.

Networking for Accelerated Computing

What to coverExplain the significance of InfiniBand NDR 400Gb/s and RoCEv2 in modern supercomputing environments. Discuss how you design network topologies to minimize latency and maximize throughput for large-scale model training.

Behavioural practice

Handling Critical Infrastructure Failure

What to coverDescribe a time you had to troubleshoot a high-priority outage in a production cluster. Focus on your diagnostic process, how you communicated the impact to stakeholders, and the steps you took to restore service and prevent recurrence.

Managing Technical Debt vs. Innovation

What to coverProvide an example of how you balanced the need to maintain legacy HPC systems with the requirement to implement new AI-driven infrastructure. Explain how you prioritized tasks and managed expectations with your team or management.

Cross-Functional Collaboration

What to coverDescribe a situation where you had to explain complex infrastructure requirements to non-technical stakeholders or data scientists. How did you ensure alignment between their workload needs and the technical constraints of the platform?

Adapting to Rapid Technological Change

What to coverGive an example of how you stayed current with emerging technologies like Agentic AI or new GPU architectures. Explain how you applied this knowledge to improve an existing process or solve a specific infrastructure challenge.

Sharpen your knowledge

Topics to prepare

HPC and AI Cluster Architecture
Review the design principles for large-scale GPU clusters, specifically focusing on multi-rail InfiniBand and RoCEv2 networking topologies. This is critical for the role's requirement to design supercomputing environments and manage 1,000+ GPU clusters.
Hybrid Cloud HPC Strategy
Study the integration of on-premises HPC resources with public cloud providers like Azure or AWS. Focus on burst strategies, FinOps for reserved/spot instances, and maintaining low-latency connectivity between hybrid environments.
Agentic AI for Infrastructure Automation
Explore the conceptual application of Agentic AI in workflow automation and observability. This is a preferred qualification for automating infrastructure management tasks and improving operational efficiency.

Skills to revise

HPC Cluster Orchestration (Slurm)
Practice configuring and managing Slurm for job scheduling and resource allocation in a multi-user GPU environment. This is a core requirement for managing XPU-based clusters and providing technical support for distributed training.
Infrastructure-as-Code (Terraform/Ansible)
Develop and test automation scripts for deploying HPC infrastructure components. Focus on creating reproducible configuration baselines and automated deployment pipelines for containerized platforms.
Advanced CUDA and GPU Management
Perform hands-on exercises with NVIDIA DGX SuperPOD architectures and CUDA optimization. This is essential for the role's focus on managing high-performance AI workloads and troubleshooting distributed training performance.

Your preparation checklist

Prepare a technical case study on large-scale GPU cluster management
Draft a detailed walkthrough of a project where you managed a large-scale GPU cluster (ideally 1,000+ units). Focus on your specific role in designing the architecture, handling InfiniBand/RoCEv2 networking, and managing orchestration tools like Slurm or Kubernetes to ensure high availability.
Demonstrate Infrastructure-as-Code (IaC) proficiency
Prepare code samples or a repository demonstrating your use of Terraform and Ansible to automate HPC environment deployment. Highlight how you have used these tools to manage configuration baselines and ensure consistency across hybrid cloud environments.
Articulate your experience with AI/HPC integration
Be ready to explain how you have integrated MLOps pipelines (e.g., Kubeflow, MLflow) with underlying HPC infrastructure. Provide examples of how you optimized distributed training workloads for frameworks like PyTorch or TensorFlow on NVIDIA H100/A100 hardware.
Develop strategic questions regarding infrastructure scalability
Prepare thoughtful questions for the interviewers about their current approach to hybrid cloud bursting, their strategy for managing energy efficiency in large-scale supercomputing environments, and how they are currently incorporating Agentic AI into their infrastructure observability workflows.

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 India. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…