New opportunity

Vertex AI & GCP Architect

Capgemini · Pune, India

About this role

Capgemini is seeking an experienced Vertex AI Engineer and Cloud Delivery Lead based in Pune, India to drive the design, deployment, and operationalisation of machine learning solutions on Google Cloud. This role bridges AI/ML engineering and cloud delivery, ensuring that models and pipelines reach production reliably, securely, and at scale.

Responsibilities

Design and implement end-to-end MLOps pipelines on Vertex AI, including data ingestion, model training, evaluation, and deployment
Build and manage Vertex AI Pipelines (Kubeflow Pipelines) for automated model training and retraining workflows
Deploy and manage models using Vertex AI Model Registry, Endpoints, and Batch Prediction services
Implement feature engineering workflows using Vertex AI Feature Store
Develop GCP-native integrations connecting Vertex AI with BigQuery, Dataflow, Cloud Storage, and Pub/Sub
Manage infrastructure for ML workloads using Terraform, ensuring reproducible and version-controlled environments
Configure IAM policies for Vertex AI workloads including service account governance and VPC Service Controls
Lead cloud delivery activities: sprint planning, release management, environment promotion, and stakeholder communication
Establish model monitoring using Vertex AI Model Monitoring for data drift and skew detection
Collaborate with data scientists to containerise experiments and promote models through dev/staging/production
Drive adoption of GKE for model serving workloads where custom inference infrastructure is required

Required qualifications

Total years of experience min. 12 years
4+ years of experience with GCP, including 2+ years hands-on with Vertex AI
Strong proficiency in Python and ML frameworks (TensorFlow, PyTorch, Scikit-learn)
Experience building Vertex AI Pipelines and managing model lifecycle in Vertex AI Model Registry
Solid Terraform skills for provisioning Vertex AI, GCS, BigQuery, and associated infrastructure
Good understanding of GCP IAM, particularly for securing ML pipelines and data access
Experience with GCP-native development patterns and event-driven architectures
Demonstrated cloud delivery experience including planning, execution, and stakeholder management
Familiarity with containerisation (Docker) and GKE for model serving

Benefits

Range of career paths and internal opportunities within Capgemini group with personalized career guidance
Comprehensive wellness benefits including health checks, telemedicine, insurance with top-ups, elder care, partner coverage, or new parent support via flexible work

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

Skills for this role

Vertex AIGoogle CloudMLOpsKubeflow PipelinesVertex AI PipelinesVertex AI Model RegistryVertex AI Feature StoreBigQueryDataflowCloud StoragePub/SubTerraformIAMVPC Service ControlsGKEPythonTensorFlowPyTorchScikit-learnDocker

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 MLOps Pipeline Architecture on Google Cloud
Design and implement scalable MLOps pipelines using Vertex AI, connecting services like BigQuery, Dataflow, Cloud Storage, and Pub/Sub to automate model training, evaluation, and deployment workflows.
Infrastructure Provisioning and Security Governance with Terraform
Manage reproducible and version-controlled environments for machine learning workloads, configuring complex GCP IAM policies, service account governance, and VPC Service Controls.
Cloud Delivery Leadership and Stakeholder Alignment
Lead sprint planning, release management, environment promotion, and stakeholder communication to successfully bridge technical AI/ML engineering teams with cloud delivery objectives.
Model Serving and Lifecycle Management via GKE and Vertex AI
Deploy, monitor, and manage machine learning models through Vertex AI Model Registry, Endpoints, Batch Prediction, and Google Kubernetes Engine for custom serving infrastructure while tracking data drift.

Likely challenges to think through

Balancing Complex Security Constraints with High-Performance ML Pipelines
Navigating stringent GCP IAM policies, service account governance, and VPC Service Controls without introducing friction into automated data ingestion and model retraining cycles. Prepare by reviewing best practices for secure multi-tenant GCP architectures and least-privilege service account configurations.
Bridging Data Science Experimentation with Production-Grade Engineering
Translating loose data science notebooks and models into robust, containerized artifacts (Docker, GKE) that adhere to rigorous environment promotion standards (dev, staging, production). Prepare by examining frameworks for standardizing model packaging and dependency management across diverse ML frameworks like TensorFlow and PyTorch.

Practice your answers

Technical practice

Designing End-to-End MLOps Pipelines on Vertex AI

What to coverDiscuss how you would architect an end-to-end MLOps pipeline on Google Cloud using Vertex AI, connecting services like BigQuery and Dataflow for data ingestion. Address how you manage data validation, model training, and automated deployment while ensuring reproducibility.

Building Automated Workflows with Vertex AI Pipelines

What to coverExplain your approach to constructing and managing Vertex AI Pipelines (Kubeflow Pipelines) for automated model training and retraining. Address component modularity, passing artifacts between steps, and handling schedule-based or event-driven triggers.

Model Registry and Inference Serving Strategies

What to coverDescribe how you leverage the Vertex AI Model Registry alongside Endpoints and Batch Prediction services for model lifecycle management. Compare when to use standard Vertex AI endpoints versus deploying custom inference workloads on Google Kubernetes Engine (GKE).

Feature Engineering and Management with Vertex AI Feature Store

What to coverExplain how you implement feature engineering workflows using Vertex AI Feature Store to serve consistent features for both training and online/offline inference. Discuss strategies for handling feature skew and point-in-time correctness.

GCP-Native Event-Driven Integrations for ML Workloads

What to coverDiscuss how you integrate Vertex AI with other GCP services such as Pub/Sub, Cloud Storage, and Dataflow in an event-driven architecture. Explain how incoming streaming data triggers automated pipeline runs or batch predictions.

Infrastructure as Code with Terraform for ML Workloads

What to coverDescribe your experience using Terraform to provision and manage reproducible environments for Vertex AI, GCS, and BigQuery. Discuss how you handle state management, module reuse, and version control across multiple deployment stages.

Security, IAM Governance, and VPC Service Controls on GCP

What to coverExplain how you configure IAM policies, service account governance, and VPC Service Controls to secure machine learning pipelines and sensitive data access on Google Cloud. Address principle-of-least-privilege implementation for automated jobs.

Model Monitoring, Data Drift, and Skew Detection

What to coverDiscuss how you configure Vertex AI Model Monitoring to detect data drift and training-serving skew in production models. Outline your strategy for establishing alert thresholds and triggering automated retraining workflows upon drift detection.

Containerization and Model Promotion Across Environments

What to coverExplain your workflow for collaborating with data scientists to containerize ML experiments using Docker and promote models through dev, staging, and production environments. Address dependency management across frameworks like TensorFlow and PyTorch.

Behavioural practice

Leading Cloud Delivery and Managing Stakeholders

What to coverDescribe a situation where you led cloud delivery activities, such as sprint planning or release management, for a complex technical initiative. Explain how you aligned cross-functional stakeholders and managed competing priorities using the Situation-Action-Result framework.

Bridging Data Science and Production Engineering

What to coverShare an example of when you had to take an experimental machine learning prototype from data scientists and productionize it. Discuss how you resolved friction between experimental flexibility and enterprise production standards.

Troubleshooting Production Pipeline Failures Under Pressure

What to coverProvide an instance where a critical MLOps pipeline failed in a production environment. Explain your diagnostic process, how you leveraged monitoring tools to identify the root cause, and the steps you took to prevent recurrence.

Navigating Strict Security and Governance Requirements

What to coverDiscuss a time when you had to implement rigorous security controls or IAM policies within a cloud environment without stalling development velocity. Highlight how you communicated trade-offs and collaborated with security teams.

Sharpen your knowledge

Topics to prepare

End-to-End MLOps Architecture on Google Cloud
Review conceptual frameworks for orchestrating machine learning workflows from data ingestion to automated model training, evaluation, and deployment. Exercise: Draft an architectural blueprint outlining how Vertex AI interacts with BigQuery, Dataflow, and Cloud Storage for an event-driven ML pipeline. Relevance: Directly addresses the core requirement to drive the design and operationalisation of machine learning solutions on GCP.
Model Monitoring, Governance, and Security on GCP
Study cloud security best practices, including GCP IAM policies, service account governance, VPC Service Controls, and automated data drift/skew detection. Exercise: Map out a secure access control matrix and monitoring strategy using Vertex AI Model Monitoring. Relevance: Aligns with responsibilities for configuring IAM policies and establishing robust model monitoring for production workloads.
Cloud Delivery and Stakeholder Management Methodologies
Examine enterprise cloud delivery processes covering sprint planning, release management, and environment promotion across dev, staging, and production. Exercise: Outline a release management lifecycle plan for promoting containerized ML models. Relevance: Essential for leading cloud delivery activities and collaborating effectively with cross-functional stakeholders.

Skills to revise

Vertex AI Pipelines and Model Registry
Practice building automated model training and retraining workflows using Vertex AI Pipelines (Kubeflow Pipelines) and managing the model lifecycle through Vertex AI Model Registry. Exercise: Write a Python script using the Vertex AI SDK to construct and execute a pipeline component that trains and registers a model. Relevance: Explicitly required for managing automated workflows and model lifecycles.
Terraform for Infrastructure as Code on GCP
Build proficiency in provisioning and managing reproducible infrastructure for machine learning workloads. Exercise: Write Terraform configuration scripts to provision a Vertex AI environment alongside Google Cloud Storage and BigQuery datasets. Relevance: Required for managing infrastructure reproducibility and version-controlled environments.
Containerization and Model Serving with GKE and Docker
Gain hands-on practice containerizing machine learning experiments and deploying custom inference workloads using Docker and Google Kubernetes Engine. Exercise: Create a Dockerfile to package a Scikit-learn or TensorFlow model and configure a deployment manifest for GKE. Relevance: Matches requirements for driving the adoption of GKE for custom model serving inference infrastructure.

Your preparation checklist

Prepare an End-to-End MLOps Architecture Walkthrough
Draft a comprehensive project presentation or diagram showcasing a complete machine learning pipeline on Google Cloud. Completion looks like having a ready-to-present architecture detailing how BigQuery, Dataflow, and Vertex AI interact for automated data ingestion, model training, and deployment.
Build a Sample Vertex AI Pipeline and Model Registry Workflow
Develop a hands-on demonstration using Vertex AI Pipelines (Kubeflow Pipelines) and Python to automate a model training workflow and register artifacts in the Vertex AI Model Registry. Completion looks like a working code repository with reproducible pipeline runs.
Create a Terraform Infrastructure Provisioning Template
Write and test Terraform configuration scripts to provision a reproducible GCP environment including Vertex AI, Google Cloud Storage, and BigQuery datasets. Completion looks like a validated set of version-controlled .tf files adhering to infrastructure-as-code best practices.
Document IAM Security and Governance Strategies
Prepare notes and architectural patterns detailing how to configure GCP IAM policies, service account governance, and VPC Service Controls for secure ML workloads. Completion looks like a clear reference guide explaining least-privilege access and data security measures.
Outline a Cloud Delivery and Stakeholder Management Plan
Synthesize past experience leading technical teams through sprint planning, release management, and environment promotion across dev, staging, and production. Completion looks like a structured summary of methodologies used to align cross-functional engineering and data science stakeholders.
Formulate Targeted Questions for the Employer
Develop thoughtful, role-specific questions regarding enterprise MLOps adoption, GKE custom inference scaling, and cross-functional team structures at Capgemini. Completion looks like a curated list of 4-5 insightful questions to ask during the interview process.

Availability last checked: 4 Nov 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…