New opportunity

AI engineer - Advisor

NTT Data · Bangalore, India

About this role

The AI Engineer - Advisor at NTT DATA in Bangalore is responsible for designing, developing, and integrating advanced AI and Generative AI applications. The role focuses on leveraging Large Language Models (LLMs) and modern AI frameworks to build scalable enterprise solutions. Responsibilities include: AI & Generative AI Development: Designing and integrating AI agents and agentic workflows for reasoning and task execution; implementing RAG solutions using vector databases and semantic search; and developing reusable AI components and APIs. LLM & Agentic AI Engineering: Integrating commercial and open-source LLMs; developing multi-agent orchestration; implementing guardrails and validation mechanisms; and optimizing model performance for quality and cost. RAG & Knowledge Engineering: Building ingestion pipelines for structured and unstructured data; implementing document parsing and chunking; and optimizing retrieval quality. Integration & API Development: Developing REST APIs and microservices to integrate AI with enterprise systems, SaaS platforms, and databases. Testing & AI Evaluation: Building evaluation datasets and automated tests to ensure accuracy, safety, and reliability of AI outputs. Deployment & LLMOps/MLOps: Implementing CI/CD pipelines, containerizing applications with Docker/Kubernetes, and managing model lifecycles. Security & Responsible AI: Implementing security controls to protect sensitive data and ensuring transparency and auditability. Required Qualifications: Strong programming skills in Python; hands-on experience with Generative AI, LLMs, RAG, and Agentic AI; experience with LLM platforms (OpenAI, Azure OpenAI, Anthropic, Google Gemini, AWS Bedrock); proficiency in AI frameworks (LangChain, LangGraph, Semantic Kernel, AutoGen); experience with vector databases (Pinecone, FAISS, Milvus, etc.); and knowledge of cloud-native development (Azure, AWS, GCP), Docker, Kubernetes, and CI/CD. Preferred Qualifications: Experience with multi-agent solutions, conversational AI, MLOps/LLMOps practices, and integrating AI with enterprise platforms like ServiceNow, Salesforce, or SAP.

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

Skills for this role

PythonGenerative AILLMsRAGAgentic AIOpenAIAzure OpenAIAnthropicGoogle GeminiAWS BedrockLangChainLangGraphSemantic KernelAutoGenAzure AI SearchPineconeElasticsearchOpenSearchpgvectorFAISSMilvusREST APIsJSONMicroservicesSQLNoSQLGitDockerKubernetesCI/CD

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

Agentic Workflow Development
The role requires designing and orchestrating multi-agent systems capable of reasoning, tool calling, and task execution. You must demonstrate proficiency in frameworks like LangGraph or AutoGen to build complex, stateful workflows that integrate with existing enterprise systems.
Production-Grade RAG Implementation
You are expected to build end-to-end RAG pipelines, from document parsing and chunking to vector database integration and retrieval optimization. Success depends on your ability to improve grounding and reduce hallucinations through techniques like reranking and hybrid search.
LLMOps and Deployment Lifecycle
The position demands a strong focus on the operational side of AI, including CI/CD for model deployment, containerization with Docker/Kubernetes, and implementing observability for monitoring latency, cost, and token usage in production environments.
Enterprise Integration and Security
You must bridge the gap between AI models and enterprise platforms (e.g., Salesforce, SAP). This involves developing secure REST APIs, implementing robust authentication, and applying guardrails to prevent data leakage and prompt injection.

Likely challenges to think through

Balancing Model Performance and Cost
A significant challenge is optimizing LLM interactions for both quality and cost. Prepare to discuss how you evaluate model performance, select appropriate LLM providers, and implement caching or parameter tuning to maintain efficiency at scale.
Ensuring AI Reliability and Safety
Implementing guardrails and validation mechanisms is critical. You will face the difficulty of ensuring AI outputs are accurate and safe in a corporate environment. Be ready to explain your approach to building automated evaluation datasets and testing for edge cases.
Managing Complex Multi-Agent State
Orchestrating multiple agents introduces risks regarding state management and error propagation. Prepare to discuss how you handle memory, state persistence, and human-in-the-loop interventions to ensure reliable task completion.

Practice your answers

Technical practice

Designing Agentic Workflows

What to coverExplain how you would architect a multi-agent system using frameworks like LangGraph or AutoGen. Focus on how you manage state, tool-calling capabilities, and the hand-off logic between agents to ensure reliable task execution.

Optimizing RAG Pipelines

What to coverDescribe your approach to improving retrieval quality in a RAG system. Discuss techniques such as document chunking strategies, hybrid search, reranking, and how you handle metadata enrichment to reduce hallucinations.

LLM Guardrails and Validation

What to coverHow do you implement input and output validation to prevent prompt injection and ensure AI outputs meet enterprise safety standards? Suggest specific mechanisms for monitoring and filtering model responses.

Vector Database Selection and Performance

What to coverCompare different vector databases like Pinecone, FAISS, or Milvus in the context of enterprise scalability. Explain how you would optimize index performance and query latency for large-scale datasets.

LLMOps and Model Lifecycle

What to coverDetail your strategy for managing the lifecycle of an LLM-based application, including versioning prompts, monitoring token usage, and tracking model performance in production using CI/CD pipelines.

Integrating AI with Enterprise Systems

What to coverHow do you securely integrate AI agents with existing enterprise platforms like Salesforce or SAP? Address authentication, secrets management, and the design of REST APIs to facilitate these connections.

Evaluating AI Application Performance

What to coverWhat metrics and methodologies do you use to evaluate the accuracy and groundedness of LLM outputs? Suggest how to build automated evaluation datasets to test for consistency and reliability.

Cloud-Native Deployment

What to coverExplain your process for containerizing an AI application using Docker and deploying it to a Kubernetes cluster. Discuss how you handle environment-specific configurations and scaling requirements.

Behavioural practice

Handling Technical Ambiguity

What to coverDescribe a situation where you had to design an AI solution with unclear or evolving requirements. How did you collaborate with stakeholders to define the scope and ensure the final architecture met business needs?

Cross-functional Collaboration

What to coverProvide an example of working with data, infrastructure, or cybersecurity teams to deploy a complex AI feature. How did you manage dependencies and resolve conflicting technical priorities?

Managing AI Risks and Ethics

What to coverTell me about a time you identified a potential risk—such as data leakage or bias—in an AI project. What actions did you take to mitigate the risk while maintaining the project's functionality?

Learning and Adapting to New AI Frameworks

What to coverThe AI field evolves rapidly; describe a time you had to quickly learn and implement a new framework or tool to solve a specific project challenge. What was your process for mastering the technology?

Sharpen your knowledge

Topics to prepare

Agentic AI Orchestration
The role emphasizes designing multi-agent workflows for reasoning and task execution. Review concepts of agent memory, state management, and tool-calling mechanisms. Exercise: Design a conceptual architecture for a multi-agent system where one agent performs information retrieval and another performs data synthesis, documenting how they share state.
RAG Pipeline Optimization
Focus on the end-to-end RAG lifecycle, including document parsing, chunking strategies, and retrieval quality improvements like reranking. Exercise: Compare different chunking strategies (e.g., fixed-size vs. semantic) for a sample technical document and explain how each affects retrieval accuracy in a vector search context.
LLMOps and Evaluation
The position requires building evaluation datasets and implementing CI/CD for AI. Review methods for measuring LLM output reliability and safety. Exercise: Create a small test suite using a framework to evaluate a prompt's performance against a set of ground-truth answers, focusing on metrics like groundedness and relevance.

Skills to revise

LangChain/LangGraph Implementation
Proficiency in these frameworks is explicitly required for building agentic workflows. Exercise: Build a simple Python script using LangGraph to create a stateful agent that can call a mock tool to fetch data and return a structured JSON response.
Vector Database Integration
Hands-on experience with databases like Pinecone, FAISS, or Milvus is required. Exercise: Set up a local FAISS index, perform a semantic search query on a small dataset of text chunks, and implement a basic filtering mechanism to refine results.
Containerization and CI/CD
The role involves deploying AI applications using Docker and Kubernetes. Exercise: Write a Dockerfile for a Python-based REST API that serves an LLM-powered endpoint, and outline a basic CI/CD pipeline step that would run automated tests on this container.

Your preparation checklist

Demonstrate RAG and Agentic Workflow Proficiency
Prepare a technical walkthrough of a project where you implemented a RAG pipeline using a vector database (e.g., Pinecone or FAISS) and an orchestration framework like LangChain or LangGraph. Focus on how you handled document chunking, semantic search, and agentic tool-calling to solve a specific enterprise task.
Showcase LLMOps and Deployment Expertise
Be ready to discuss your experience with the full AI lifecycle, specifically how you containerized an AI application using Docker and deployed it to Kubernetes. Highlight your approach to CI/CD pipelines, monitoring model performance, and managing costs or latency in a production environment.
Prepare for Enterprise Integration Discussions
Review your experience integrating AI models with enterprise systems like Salesforce, SAP, or ServiceNow. Prepare to explain how you secured these integrations using REST APIs, authentication protocols, and secrets management to protect sensitive data.
Formulate Strategic Questions for the Interviewer
Prepare thoughtful questions regarding the team's current approach to AI evaluation and guardrails. Ask about how they balance model accuracy with safety requirements or how they manage multi-agent orchestration in their existing enterprise architecture.

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…