New opportunity

Technology Platform Engineer

Accenture · Bengaluru, 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 production and non-production cloud environments, deploying automation pipelines, and managing XPU-based clusters (CPU/GPU/accelerators) to provide Metal as a Service (MaaS), GPUaaS, and AIaaS offerings. Responsibilities include optimizing cluster performance, scalability, and cost efficiency across on-premises, cloud, and hybrid environments, as well as integrating AI/HPC platforms with existing IT systems and security frameworks. The engineer will monitor, troubleshoot, and tune infrastructure to ensure high availability and low-latency networking, while providing technical guidance to users for large-scale models and simulations. Required qualifications include at least 5 years of experience in enterprise-wide HPC strategy and architecting supercomputing environments, with deep knowledge of accelerated computing architectures, high-performance fabrics, and cluster orchestration. Candidates must possess hands-on experience with Linux, advanced CUDA/GPU management, and infrastructure-as-code tools like Terraform and Ansible. Preferred qualifications include experience managing 1,000+ GPU clusters, familiarity with AI frameworks (TensorFlow, PyTorch, JAX), and knowledge of Agentic AI technologies for workflow automation. A 15-year full-time educational background is required.

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

Skills for this role

Linux ArchitectureHPC cluster designCUDAGPUSLURMPBS ProMPIOpenMPSYCLNVIDIA DGX SuperPODInfiniBandRoCEv2MLOpsKubeflowSingularityKubernetesAnsibleTerraformAzure MLMLflowPythonMachine LearningCloud Technology ArchitectureDockerEdge ComputingMicrosoft Agentic AI ArchitectTensorFlowPyTorchJAX

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 Scalable XPU-Based Clusters
The role requires designing and managing high-performance computing environments, specifically focusing on GPUaaS and MaaS offerings. You must demonstrate expertise in orchestrating large-scale clusters (1,000+ GPUs) using tools like Slurm and Kubernetes, while ensuring high-performance fabric connectivity such as InfiniBand or RoCEv2.
Infrastructure Automation and IaC
You are expected to deploy and maintain production and non-production environments through automation pipelines. Proficiency in Infrastructure-as-Code (IaC) using Terraform and Ansible is critical to ensure consistent configuration, environment reproducibility, and efficient cluster management across hybrid cloud setups.
Performance Tuning and Observability
A core responsibility is monitoring and optimizing infrastructure for low-latency networking and high availability. You must be prepared to troubleshoot complex bottlenecks in AI/HPC workloads and implement performance tuning strategies to maximize cluster utilization and cost efficiency.

Likely challenges to think through

Balancing Performance with Cost Efficiency
Managing large-scale GPU clusters involves significant infrastructure costs. The challenge lies in implementing FinOps strategies, such as utilizing Spot VMs or Reserved Instances, without compromising the high-performance requirements of AI and HPC workloads. Prepare to discuss how you balance technical performance needs with budgetary constraints.
Integrating Heterogeneous Environments
The role requires bridging on-premises supercomputing environments with cloud-based AI platforms. This hybrid integration often involves complex networking and security frameworks. Prepare to explain how you manage data gravity, latency, and security consistency when workloads burst between local clusters and public cloud providers.

Practice your answers

Technical practice

Designing High-Performance Fabrics

What to coverDiscuss the trade-offs between InfiniBand and RoCEv2 in a large-scale GPU cluster environment. Focus on latency, throughput, and the complexity of configuring multi-rail designs for H100/A100 clusters.

HPC Cluster Orchestration and Scheduling

What to coverExplain how you would manage resource contention in a multi-tenant environment using Slurm or Kubernetes. Address how you balance job priority, preemption, and fair-share policies to maximize GPU utilization.

Infrastructure as Code for HPC

What to coverDescribe your approach to using Terraform or Ansible to automate the deployment of reproducible HPC environments. Suggestion: Focus on how you handle state management and configuration drift in hybrid cloud setups.

GPU Cluster Performance Tuning

What to coverWhat metrics and observability tools do you prioritize when troubleshooting low-latency networking or GPU underutilization? Discuss the role of NVMe-oF and parallel file systems like Lustre or Weka in preventing I/O bottlenecks.

Hybrid Cloud HPC Strategy

What to coverHow do you design a bursting strategy that moves workloads from on-premises to public cloud (e.g., Azure) while maintaining data consistency? Address cost-efficiency strategies like Spot VM usage and Reserved Instances.

Containerization in AI Workloads

What to coverCompare the use of Singularity versus Docker/Kubernetes for running large-scale AI models. Explain how you handle security and performance isolation for containerized GPU tasks.

Managing Large-Scale GPU Clusters

What to coverWhat are the primary challenges in managing a 1,000+ GPU cluster, specifically regarding power, cooling, and hardware failure rates? Suggestion: Discuss proactive monitoring and the use of NVIDIA BCM or similar management suites.

AI Framework Integration

What to coverHow do you support data scientists using frameworks like PyTorch or JAX on your infrastructure? Explain the role of MLOps tools like Kubeflow or MLflow in streamlining the transition from development to production.

Behavioural practice

Handling Critical Infrastructure Outages

What to coverDescribe a time you had to troubleshoot a high-priority production system failure under pressure. Focus on your diagnostic process, how you communicated with stakeholders, and the steps taken to prevent recurrence.

Providing Technical Guidance to Users

What to coverGive an example of when you had to explain complex infrastructure limitations to a non-infrastructure team (e.g., Data Scientists). How did you balance their performance requirements with system stability?

Managing Conflicting Priorities

What to coverDescribe a situation where you had to balance long-term architectural improvements with urgent operational support requests. How did you decide which tasks to prioritize?

Collaborating Across Teams

What to coverShare an experience where you had to integrate a new security framework or IT policy into an existing HPC environment. How did you manage the collaboration with security teams while ensuring minimal impact on workload performance?

Sharpen your knowledge

Topics to prepare

HPC and AI Infrastructure Architecture
Review the design principles for large-scale XPU-based clusters, focusing on the integration of GPUaaS and MaaS models. Exercise: Draft a high-level architecture diagram for a 1,000+ GPU cluster, detailing the interconnect strategy (InfiniBand/RoCEv2) and how you would balance performance with cost-efficiency in a hybrid cloud environment.
High-Performance Networking and Fabrics
Understand the requirements for low-latency networking in supercomputing environments. Exercise: Compare the architectural differences between InfiniBand NDR and RoCEv2, and explain how you would troubleshoot a multi-rail network congestion issue in a DGX SuperPOD deployment.
Cluster Orchestration and Scheduling
Focus on the coexistence of traditional HPC schedulers like Slurm with modern container orchestrators like Kubernetes. Exercise: Design a workflow where a user submits a distributed training job that requires both Slurm for resource allocation and Kubernetes for containerized runtime execution.

Skills to revise

Infrastructure-as-Code (Terraform and Ansible)
Demonstrate proficiency in automating environment provisioning. Exercise: Write a Terraform module to deploy a virtualized cluster node and an Ansible playbook to configure the necessary GPU drivers and MPI libraries on that node.
Advanced CUDA and GPU Management
Focus on the lifecycle management of NVIDIA hardware. Exercise: Describe the steps to monitor GPU health and performance metrics using NVIDIA tools, and explain how you would optimize a workload that is experiencing GPU memory bottlenecks.
Linux Architecture and Performance Tuning
Deep dive into kernel-level tuning for high-performance computing. Exercise: Identify three specific Linux kernel parameters or system configurations you would adjust to minimize latency for a high-throughput AI training workload.

Your preparation checklist

Prepare an HPC Infrastructure Case Study
Develop a technical walkthrough of a large-scale cluster deployment you have managed, specifically highlighting your experience with GPU orchestration (e.g., Slurm, Kubernetes, Run:ai) and high-performance fabrics like InfiniBand or RoCEv2. Focus on how you addressed performance bottlenecks and ensured low-latency networking.
Demonstrate Infrastructure-as-Code Proficiency
Create a repository or portfolio document showcasing your use of Terraform and Ansible to automate the provisioning of hybrid cloud environments. Include examples of how you manage configuration baselines and automated deployment pipelines for AI/HPC workloads.
Articulate GPU and XPU Management Strategy
Prepare to discuss your hands-on experience with NVIDIA H100/A100 architectures and CUDA management. Be ready to explain your approach to cluster optimization, including how you handle multi-tenant GPUaaS offerings and resource scheduling for distributed training tasks.
Formulate Strategic Questions for the Interviewer
Prepare insightful questions regarding the team's current approach to hybrid cloud bursting, their specific strategies for managing 1,000+ GPU clusters, and how they integrate Agentic AI technologies into their existing infrastructure observability and automation 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…