About the role
The offshore AI Infrastructure Architect owns detailed design and delivery architecture partnering with the onshore lead architect The role translates the reference architecture into codified buildable designs the IaC blueprint the tenantprovisioning framework and the nine automation pipelines and provides design authority to the offshore engineering pods Requires working overlap with US hours for design alignment
Key Responsibilities
Produce the detailed design for the AWS landing zone EKSonOutposts platform and the NVAIE inference stack from the agreed reference architecture
Own the codified singletouch sitebootstrap blueprint Rafay NVAIE RunAI as code and its evolution across sites
Design the tenantprovisioning framework and the nine automation pipelines site bootstrap tenant creation model onboarding inferencing training runtime promotionGitOps tenant lifecycle observabilityFinOps export DROutpost recovery
Detail the multitenant isolation model vCluster RBAC pertenant KMS network segmentation RunAI projectquota
Detail orchestration boundaries and integration points between Rafay RunAI and KEDA
Detail HADR mechanisms warmstandby failover crossregion EKS resiliency secretscertconfig replication and Outpost recovery flows
Provide design authority and technical review across the offshore engineering pods
Maintain the architecture riskassumptions log jointly with the onshore architect
MustHave Skills Experience
12 years in IT with 5 years in cloudplatform architecture on AWS at production scale
Deep AWS Outposts EKS VPCnetworking Direct Connect Transit Gateway KMS IAM and multiaccount landing zone Control Tower
Strong Kubernetes architecture including multitenancy and handson understanding of EKSonOutposts constraints
GPU AI infrastructure architecture NVIDIA GPUs GPU sharing MIG vs fractional GPU and GPU scheduling
HADR architecture warm standby multiregion resiliency defined RTORPO tiers
InfrastructureasCode Terraform and GitOps operating models
Strong customerhandling skills able to run architecture workshops and design reviews with the Customer and AWS present and defend design decisions to senior stakeholders manage expectations and translate business requirements into technical solutions
Stakeholder management and executive communication building trust with the Customer handling escalations and driving consensus across Customer AWS and LTM teams
AWS Certified Solutions Architect Professional or equivalent depth
Proven experience delivering detailed designs and blueprints in a distributed onshoreoffshore model
NicetoHave Skills
Rafay or Spectro Cloud RunAI and the NVIDIA Dynamo NIM NVAIE stack
Telco edge 5G MEC platforms
SageMaker training runtime data lake Lake Formation Glue
vCluster and policyascode OPA Kyverno