New opportunity

Manager, AI/HPC Infrastructure Technical Delivery — India

NVIDIA · Pune, India

About this role

NVIDIA is seeking a Senior Technical Delivery Manager (People Manager) in Pune, India, to lead a high-performing technical team overseeing end-to-end technical control for InfiniBand, HPC, and AI cluster delivery across India and APAC. The role involves guiding Senior Solutions Architects and specialists to design, deploy, and ensure operational excellence for large-scale GPU-accelerated and data center infrastructure.

Responsibilities

Lead, manage, and develop a technical team of Senior Networking Solutions Architects and specialists across the full people management lifecycle, including hiring, coaching, and career development.
Provide overall technical control for InfiniBand, HPC, and AI cluster delivery projects across the India and APAC region, adhering to NVIDIA product standards.
Oversee the design, deployment, and operational reliability of large-scale AI/HPC infrastructure using InfiniBand networking and AI orchestration tools like Base Command Manager.
Lead technical decision-making and resolve critical challenges across data center architecture, storage, servers, network automation, and GPU-accelerated systems.
Optimize delivery workflows for provisioning, system design, automation, and post-deployment monitoring, logging, and alerting.
Partner with engineering, sales, product, and support teams to align delivery strategies with NVIDIA's product roadmap.
Build trusted relationships with key customers and partners across India and APAC as the technical leadership point of contact.
Maintain technical best practices, compliance, and quality requirements while driving project delivery timelines and customer satisfaction.
Track project progress and communicate technical delivery status and solutions to senior leadership.

Required qualifications

BS, MS, PhD, or equivalent professional experience in Computer Science, Electrical/Computer Engineering, Physics, Mathematics, or a related technical field.
10+ years of progressive technical experience in data center infrastructure, or equivalent experience.
5+ years of people management experience leading a technical team of 10 or more members.
Strong technical background in NVIDIA DC products, InfiniBand, AI orchestration tools (e.g., Base Command Manager), GPU-accelerated infrastructure, and full-stack data center technologies.
Deep understanding of AI technology and its integration with data center infrastructure.
Deep expertise in networking fundamentals, TCP/IP stack, data center architecture, and hands-on experience in medium to large-scale HPC/AI environments.
Demonstrated experience managing end-to-end technical delivery of large-scale AI/HPC infrastructure projects.
Proven track record resolving complex technical challenges and optimizing delivery processes.
Excellent customer-facing and bilingual communication skills, with a solid understanding of the India business model and ability to collaborate across time zones.

Preferred qualifications

Linux or Networking Certifications (e.g., CCNP, CCIE, HCIE) or NVIDIA-related certifications.
Extensive experience with HPC job schedulers such as Slurm or PBS, and cluster management technologies like Base Command Manager.
Experience leading technical teams in a professional services or tech vendor environment working with top CSP, cloud, or hyper-scale customers in India.
Deep understanding of AI/ML workloads and hands-on experience optimizing performance for GPU-accelerated systems.

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

Skills for this role

AI InfrastructureHPCInfiniBandEthernetBase Command ManagerGPU AccelerationData Center ArchitectureNetwork AutomationVirtualizationTCP/IPSlurmPBSLinuxPeople ManagementTechnical Delivery

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

Technical Governance for AI and HPC Cluster Deployments
Maintain end-to-end technical oversight across large-scale InfiniBand, Ethernet, and GPU-accelerated infrastructure projects. This requires enforcing NVIDIA data center standards, resolving architecture bottlenecks across compute, network, and storage, and overseeing cluster orchestration tools such as Base Command Manager.
People Leadership and Specialist Team Development
Manage the full talent lifecycle for a regional team of senior solutions architects and technical delivery specialists. Key duties include hiring, technical coaching, performance management, and setting high engineering delivery standards while supporting distributed teams across time zones.
Regional Customer and Stakeholder Relationship Management
Serve as the primary technical delivery authority for enterprise, CSP, and hyperscale accounts across India and APAC. The role demands aligning delivery scope with regional business dynamics, managing customer expectations during complex rollouts, and bridging communication with Chinese and regional partners.
Delivery Process Optimization and Operational Readiness
Standardize and automate deployment workflows—including network provisioning, cluster configuration, telemetry, monitoring, and post-deployment validation—to safeguard cluster reliability and keep delivery schedules aligned with product roadmaps.

Likely challenges to think through

Balancing High-Touch People Management with Deep Technical Escalations
Leading a team of 10+ senior architects while maintaining technical authority over complex cluster deployments risks operational bottlenecks. Prepare examples demonstrating how to delegate architecture responsibilities effectively, maintain structured escalation paths, and lead technical reviews without micromanaging senior staff.
Triaging Multi-Vendor Infrastructure Dependencies Under Aggressive Timelines
Deploying cutting-edge AI clusters requires integrating NVIDIA DC hardware, InfiniBand fabric, custom storage solutions, and third-party facility constraints. Prepare case studies illustrating how you diagnose fabric-level performance issues, isolate root causes across hardware versus software, and renegotiate critical-path milestones under pressure.
Managing Regional Business Nuances and Cross-Border Customer Expectations
Coordinating technical delivery across India and APAC involves navigating diverse customer maturity levels, distinct regional business practices, and multilingual communication requirements. Prepare to explain your approach to managing cross-cultural client engagements, standardizing SLAs across borders, and handling complex customer escalations.

Practice your answers

Technical practice

InfiniBand vs. RoCE Architectural Design in AI Clusters

What to coverExplain the architectural trade-offs between dedicated InfiniBand fabrics and RDMA over Converged Ethernet (RoCE) for large-scale distributed AI workloads. Detail how you evaluate congestion management, latency profiles, subnet administration, and operational overhead when establishing technical delivery standards.

Cluster Provisioning and Lifecycle Management with Base Command Manager

What to coverWalk through the end-to-end bare-metal provisioning and verification lifecycle of a GPU compute cluster using NVIDIA Base Command Manager (BCM). Highlight how you configure OS deployment, driver validation, network topology mapping, and automated health checks to ensure initial delivery readiness.

GPU Interconnect and Fabric Topology Bottleneck Triage

What to coverDescribe how you lead your team through diagnosing performance degradation during large-scale all-reduce collective communications across multiple GPU nodes. Address the diagnostic approach across NVLink, PCIe subsystems, host channel adapters (HCAs), optical transceivers, and InfiniBand switch routing.

Full-Stack Data Center Storage and Compute Integration

What to coverHow do you evaluate and validate data center storage architectures to support high-throughput GPU training without starving compute nodes? Discuss the integration requirements for parallel file systems, GPUDirect Storage, and the network transport layer under peak concurrent read/write loads.

HPC Job Scheduler Optimization for AI Workloads

What to coverExplain how you configure and supervise workload orchestration using schedulers like Slurm or PBS in a shared AI/HPC environment. Cover strategies for topology-aware GPU allocation, preemption policies, and handling node health evictions during active model training runs.

Network Automation and Telemetry for Large-Scale Fabrics

What to coverOutline your approach to automating network provisioning and post-deployment telemetry (monitoring, logging, and alerting) across hundreds of fabric nodes. Detail the key performance indicators and error counters you monitor to detect transient link flapping, packet drops, or thermal throttling early.

Technical Governance and Architectural Alignment with Product Roadmaps

What to coverAs technical delivery manager, how do you bridge field implementation challenges with NVIDIA DC product engineering roadmaps? Provide the structured approach you use to feed customer architecture constraints, firmware bugs, and feature gaps back into core product and engineering teams.

Managing Regional Data Center Physical and Environmental Constraints

What to coverDiscuss the technical constraints commonly encountered when deploying high-density GPU infrastructure across regional data center facilities. Address how your team calculates and verifies power distribution limits, rack thermal dissipation, cooling redundancy, and spatial rack layouts.

Behavioural practice

Leading and Coaching a Senior Technical Team Through Complex Deliveries

What to coverDescribe a situation where you managed a team of senior technical specialists facing a stalled or high-friction infrastructure deployment. Walk through how you diagnosed the operational bottleneck, coached the engineers, set clear technical ownership, and achieved the final delivery outcome.

Navigating a High-Stakes Customer Escalation with Strategic Accounts

What to coverShare an example where an enterprise, hyperscale, or cloud service provider client experienced critical deployment delays or performance shortfalls. Detail how you served as the leadership point of contact, balanced customer business expectations against engineering reality, and rebuilt trust.

Cross-Cultural and Regional Stakeholder Alignment

What to coverRecall a project where you coordinated delivery across distinct geographic regions, time zones, or cross-border partners (such as APAC or Chinese customer engagements). Explain the specific communication or business model adjustments you implemented to drive consensus and meet project milestones.

Managing Delivery Timeline Delays Against Strict Quality Standards

What to coverProvide an example where commercial delivery deadlines pressured your team to bypass compliance or technical quality validation steps on a major deployment. Explain how you evaluated the trade-off, defended NVIDIA-grade engineering standards with stakeholders, and navigated project schedules.

Sharpen your knowledge

Topics to prepare

InfiniBand Architecture vs. Ethernet in Large-Scale AI Fabrics
Review subnet management, fat-tree topologies, adaptive routing, congestion control, and RDMA over Converged Ethernet (RoCE) vs. native InfiniBand. This is directly relevant to the role's mandate of providing overall technical control over Ethernet, InfiniBand, and AI cluster delivery. Exercise: Map out an end-to-end fabric architecture for a multi-rack GPU cluster, contrasting failover behavior, packet loss mitigation, and latency characteristics between InfiniBand and RoCE.
Full-Stack Data Center Integration for GPU Computing
Examine the technical interplay between high-throughput parallel storage (e.g., GPUDirect Storage), server compute, virtualization, power, and high-density GPU cooling. This directly addresses the requirement for full-stack data center technologies. Exercise: Outline a design review checklist evaluating power distribution, thermal envelopes, and PCI/NVLink topology constraints for deploying high-density GPU nodes.
HPC and AI Workload Orchestration & Job Scheduling
Study workload execution dynamics, batch queue management, resource partitioning, and job scheduling mechanisms using Slurm or PBS alongside Base Command Manager (BCM). This supports the explicit preference for scheduler expertise and deployment orchestration. Exercise: Formulate a strategy for handling node health evictions, priority queuing, and multi-node distributed training checkpoints in a busy cluster.
Technical Escalation Governance and Engineering Leadership
Understand structural methods for leading a 10+ member team of senior architects, managing high-stakes escalations, and balancing client SLAs against product engineering constraints. This aligns with the requirement for 5+ years managing technical teams and customer-facing delivery. Exercise: Draft a post-incident review (RCA) framework and communication template for reporting cluster delivery delays or fabric downtime to executive leadership.

Skills to revise

Cluster Provisioning & Orchestration with Base Command Manager
Hands-on configuration of node provisioning, image distribution, and monitoring via NVIDIA Base Command Manager (BCM). The posting cites BCM as a key tool for cluster management and automation. Exercise: Simulate or document the step-by-step workflow for bare-metal discovery, OS image deployment, network interface configuration, and GPU health checks across a cluster using BCM.
Fabric Performance Diagnosis & Latency Isolation
Practical network diagnostic commands and telemetry analysis for InfiniBand and TCP/IP fabrics. The role requires resolving critical technical challenges in data center architecture and InfiniBand networks. Exercise: Write down a structured diagnostic triage path to trace performance degradation during distributed GPU collective operations (e.g., isolating bad transceivers, link errors, or mismatched MTUs).
HPC Job Scheduler Configuration (Slurm / PBS)
Configuring job schedulers for GPU-accelerated computing and multi-node workloads, matching the explicit preferred qualification for Slurm and PBS. Exercise: Write a sample Slurm configuration defining GPU resources (GRES), topology-aware node definitions, and a submission script testing inter-node GPU communication.
Technical Milestone & Customer Delivery Review Simulation
Conducting customer-facing technical delivery alignment meetings and managing project timelines across time zones. This directly exercises the requirement for customer relationship management with regional and Chinese enterprise accounts. Exercise: Role-play a delivery alignment session addressing an scope change or vendor hardware delay with a Tier-1 CSP client, presenting phased recovery milestones.

Your preparation checklist

Construct a Detailed Large-Scale AI/HPC Cluster Delivery Walkthrough
Prepare a structured case study detailing an end-to-end infrastructure rollout you led. Clearly articulate the cluster sizing, GPU topology, InfiniBand fabric design, storage integration, and operational milestones, highlighting how you met strict delivery timelines and NVIDIA-grade quality standards.
Document People Management and Engineering Leadership Evidence
Draft concrete examples demonstrating your leadership of 10+ technical professionals across hiring, career development, performance coaching, and retention. Show how you established engineering best practices and maintained high team morale in a high-demand delivery environment.
Complete an InfiniBand Fabric and Network Architecture Design Exercise
Diagram an enterprise-scale AI cluster network illustrating leaf-spine topology, subnet manager placement, routing algorithms, congestion management, and monitoring points. Contrast this with RoCE and explain the trade-offs in multi-node collective communication performance.
Outline an Orchestration and Automation Workflow Using Base Command Manager
Map out a standard operational procedure for bare-metal cluster provisioning, automated firmware updates, OS image deployment, and health validation using cluster management tools like Base Command Manager (BCM) and job schedulers (Slurm or PBS).
Synthesize Examples of Managing Strategic Regional Accounts and Escalations
Prepare two customer-facing scenarios focusing on enterprise, CSP, or hyperscale clients in India/APAC. Highlight how you managed commercial versus technical trade-offs, translated customer requirements into technical milestones, and addressed cross-border communication nuances.
Formulate Strategic Questions on Delivery Governance and Product Alignment
Develop 3-5 thoughtful questions for interviewers regarding how the India/APAC delivery team interfaces with NVIDIA product engineering, standard deployment toolchains, and emerging hardware release cycles across regional data centers.

Availability has not yet been verified. 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…