New opportunity

Lead Agentic AI Engineer

Accenture · Kochi, India

About this role

The Lead Agentic AI Engineer will design, build, and deploy production-grade Generative AI, Agentic AI, and Machine Learning applications, primarily leveraging Google Cloud Platform (GCP) and Vertex AI. This role focuses on creating intelligent agents capable of reasoning, retrieval, tool use, and multi-step task automation within enterprise environments.

Responsibilities

Develop AI agents for planning, reasoning, workflow orchestration, and multi-agent collaboration.
Implement RAG solutions using vector search, semantic search, and enterprise data.
Build scalable architectures integrating Vertex AI with GCP services like BigQuery, Cloud Run, and GKE.
Develop secure APIs, microservices, and connectors to integrate AI solutions with enterprise systems.
Utilize frameworks such as LangChain, LangGraph, LlamaIndex, or AutoGen.
Ensure production quality by managing accuracy, hallucination reduction, latency, and observability.
Collaborate with cross-functional teams including architects, data engineers, and DevOps.

Requirements

5-8 years of professional experience in AI/ML application development or software engineering.
Strong hands-on experience with GCP and Generative AI.
Proficiency in Vertex AI, Gemini models, and cloud-native development.
Understanding of Agentic AI concepts like tool calling, memory, and multi-agent systems.
Graduation degree.

Preferred Qualifications

Experience with AWS Bedrock, Azure AI Foundry, or Amazon SageMaker.
Knowledge of Model Context Protocol (MCP) and Agent2Agent (A2A) patterns.

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

Skills for this role

GCPGenerative AIVertex AIGemini modelsRAGAI agentsLLMMachine LearningAgentic AIPythonBackend EngineeringData EngineeringCloud-native developmentBigQueryCloud RunGKEPub/SubAPI developmentMicroservicesModel Context Protocol (MCP)Agent2Agent (A2A)LangChainLangGraphLlamaIndexSemantic KernelAutoGenCrewAIAWS BedrockAzure AI FoundryAmazon SageMaker

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

Production-Grade Agentic Architecture
The role demands moving beyond prototypes to scalable, secure, and observable AI systems. You must prioritize building robust workflows using Vertex AI and Gemini that handle multi-step reasoning, tool calling, and memory management while ensuring enterprise-level reliability and cost efficiency.
GCP Ecosystem Integration
Success depends on deep integration between AI models and the broader Google Cloud stack. You must demonstrate how to orchestrate Vertex AI with BigQuery, Cloud Run, GKE, and Pub/Sub to create functional, data-driven applications that interact securely with enterprise systems.
Advanced Agentic Patterns
You are expected to implement sophisticated patterns such as Model Context Protocol (MCP) and Agent2Agent (A2A) communication. Prioritize mastering these frameworks to enable modular, interoperable agent systems that can collaborate and perform complex task automation.

Likely challenges to think through

Balancing Agent Autonomy with Guardrails
A core challenge is reducing hallucinations and ensuring groundedness in autonomous agents. Prepare by discussing your approach to implementing observability, guardrails, and rigorous evaluation frameworks to maintain accuracy in high-stakes enterprise environments.
Managing Latency in Multi-Agent Systems
Complex multi-agent workflows often introduce significant latency. Prepare to explain how you optimize orchestration frameworks like LangGraph or AutoGen to ensure performance remains acceptable for production use cases without sacrificing reasoning quality.

Practice your answers

Technical practice

Architecting Agentic Workflows on GCP

What to coverExplain how you would design a multi-step agentic workflow using Vertex AI and Cloud Run. Discuss how you manage state, tool calling, and memory across multiple execution steps while ensuring scalability and cost-efficiency.

Implementing RAG with Vertex AI

What to coverDescribe your approach to building a production-grade RAG pipeline. Focus on how you handle document ingestion, vector search optimization, and grounding Gemini models to reduce hallucinations in enterprise data contexts.

Agent-to-Agent (A2A) Communication Patterns

What to coverHow do you implement A2A patterns for task delegation and inter-agent collaboration? Discuss the challenges of maintaining consistency and security when agents interact with each other and external enterprise systems.

Model Context Protocol (MCP) Integration

What to coverExplain the role of the Model Context Protocol in standardizing agent access to enterprise tools. How would you implement an MCP server to expose internal APIs or databases to an AI agent securely?

Observability and Evaluation in Production

What to coverWhat metrics and tools do you use to monitor the performance of LLM-powered applications? Detail your strategy for tracking latency, accuracy, and reliability, and how you implement guardrails to manage production risks.

Orchestration Framework Trade-offs

What to coverCompare frameworks like LangGraph, LlamaIndex, and AutoGen. In what scenarios would you choose one over the others for building complex, multi-agent enterprise applications?

Securing AI Microservices

What to coverHow do you ensure security when integrating AI agents with enterprise systems? Discuss authentication, authorization, and the use of Secret Manager or other GCP services to protect sensitive data during tool execution.

Optimizing LLM Latency and Cost

What to coverDescribe techniques for optimizing the performance of Gemini-based applications. Address strategies such as prompt engineering, caching, model selection, and infrastructure choices on GCP to balance speed and operational costs.

Behavioural practice

Cross-functional Collaboration

What to coverDescribe a situation where you had to align architects, data engineers, and DevOps teams to deliver a complex AI solution. How did you manage conflicting priorities or technical requirements?

Handling Production Incidents

What to coverTell me about a time an AI application failed in production or produced unexpected results. What was your process for identifying the root cause, implementing a fix, and preventing recurrence?

Managing Technical Debt in AI Projects

What to coverAI development often involves rapid experimentation. Give an example of how you balanced the need for quick prototyping with the requirement for building scalable, maintainable, and production-ready code.

Adapting to New AI Technologies

What to coverThe AI landscape evolves rapidly. Provide an example of how you have proactively learned and applied a new framework or pattern (like MCP or a new orchestration library) to improve an existing project.

Sharpen your knowledge

Topics to prepare

Agentic AI Orchestration Patterns
Focus on the conceptual design of multi-agent systems, including task delegation, inter-agent communication (A2A), and planning. Exercise: Draft a high-level architecture diagram for a multi-agent system where one agent handles data retrieval and another handles structured output generation, using the Model Context Protocol (MCP) for tool access.
Production-Grade RAG and Observability
Understand the lifecycle of RAG beyond simple retrieval, focusing on hallucination reduction, grounding, and latency management. Exercise: Define a strategy for monitoring an LLM application in production, specifically identifying metrics for retrieval quality and model accuracy, and how you would implement guardrails to mitigate hallucinations.

Skills to revise

Vertex AI and Gemini Integration
Hands-on implementation of Google Cloud's AI stack. Exercise: Build a prototype using Vertex AI SDK to invoke a Gemini model with function calling enabled, allowing the model to query a BigQuery dataset to answer a specific business question.
Agentic Framework Implementation
Proficiency in frameworks like LangGraph or AutoGen for workflow automation. Exercise: Create a simple state-machine workflow using LangGraph that manages a multi-step task, such as summarizing a document and then routing the summary to a specific API endpoint based on sentiment.
Cloud-Native Backend Engineering
Integrating AI services with GCP infrastructure. Exercise: Design a microservice architecture on Cloud Run that exposes an AI agent as a secure API, ensuring it interacts with Secret Manager for credentials and Pub/Sub for asynchronous task processing.

Your preparation checklist

Prepare a Vertex AI and Gemini project walkthrough
Develop a concise case study demonstrating a production-grade application built on GCP. Focus on your implementation of Vertex AI, Gemini models, and RAG pipelines, specifically highlighting how you managed latency, hallucination reduction, and cost efficiency in an enterprise context.
Demonstrate agentic workflow orchestration
Prepare a technical explanation of how you have utilized frameworks like LangGraph, LlamaIndex, or AutoGen to build multi-agent systems. Be ready to discuss your approach to tool calling, memory management, and task delegation patterns.
Articulate your experience with MCP and A2A patterns
Review your knowledge of the Model Context Protocol (MCP) and Agent2Agent (A2A) communication. Prepare to explain how you would implement these to standardize tool access and enable secure, scalable collaboration between autonomous agents in a complex enterprise environment.
Map your backend engineering skills to GCP services
Create a summary of how you integrate AI solutions with cloud-native infrastructure. Specifically, prepare to discuss your experience connecting Vertex AI with BigQuery, Cloud Run, GKE, and Pub/Sub to build secure, observable, and scalable microservices.
Formulate strategic questions for the interviewers
Prepare thoughtful questions regarding the team's current approach to AI observability, guardrails, and CI/CD for agentic workflows. Ask about how they balance rapid prototyping with the rigorous security and reliability requirements of their enterprise clients.

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…