New opportunity

AI Architect - LangChain/LangGraph

NetConnectGlobal · Gurgaon, India

About this role

NetConnectGlobal seeks an AI Architect (MCP) to lead the design and implementation of enterprise-scale agentic AI and LLM-powered solutions. The role combines strategic enterprise architecture with hands-on implementation and is based in Gurgaon in a hybrid arrangement. The posting states a 12-month contract duration and lists the employment type as full-time.

Responsibilities

Design MCP server architectures, AI agents and enterprise AI platforms; develop reusable agent skills, prompt libraries and RAG-based solutions.
Build enterprise context-engineering frameworks and AI governance standards. Design enterprise knowledge architectures, ontology models and AI accelerator frameworks.
Write Architecture Decision Records (ADRs), software development lifecycle (SDLC) standards and security review documentation.
Work with enterprise architecture, engineering and business teams to deliver scalable solutions and promote AI platform adoption, governance and best practices.

Requirements

At least 8 years of experience in software, solution or enterprise architecture, plus a B.E., B.Tech. or M.Tech. in Computer Science, IT, Engineering or a related field.
Mandatory expertise in MCP, agentic AI, enterprise AI architecture, context engineering, LangChain, LangGraph, the OpenAI or Claude APIs, prompt engineering, RAG, Python, TypeScript or Node.js, enterprise architecture, Azure, GitHub, Jira and Confluence.

Preferred skills include knowledge graphs and ontology design; TOGAF or the Zachman Framework; AI governance and security; semantic data architecture; internal developer platforms; and enterprise knowledge management.

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

Skills for this role

Model Context Protocol (MCP)Agentic AIEnterprise AI ArchitectureContext EngineeringLangChainLangGraphOpenAI APIClaude APIPrompt EngineeringRetrieval-Augmented Generation (RAG)PythonTypeScriptNode.jsEnterprise ArchitectureAzureGitHubJiraConfluenceKnowledge GraphsOntology DesignTOGAFZachman FrameworkAI GovernanceAI SecuritySemantic Data ArchitectureInternal Developer PlatformsEnterprise Knowledge ManagementSQLETL

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

Enterprise agent architecture
The role leads the design and implementation of MCP server architectures, AI agents and enterprise AI platforms. Prepare to explain architecture choices alongside how you would implement and validate them.
Reusable LLM solutions
The posting calls for reusable agent skills, prompt libraries and RAG-based solutions, with LangChain or LangGraph among the mandatory skills. Prepare an example that shows how reusable components fit into an end-to-end solution.
Context and knowledge architecture
Building context-engineering frameworks and designing enterprise knowledge architectures, ontology models and AI accelerators are stated duties. Prepare to distinguish what belongs in retrieved context, structured knowledge and agent behavior.
Governance and architecture records
The architect is expected to establish AI governance standards and author ADRs, SDLC standards and security review documentation. Prepare to show how a technical decision would be recorded and reviewed.
Cross-functional delivery and adoption
The role works with enterprise architecture, engineering and business teams to deliver scalable solutions and promote platform adoption. Prepare to explain how you would communicate architecture decisions to those different groups.

Likely challenges to think through

Balancing architecture with implementation
Likely preparation challenge, not a reported employer problem: enterprise-wide designs may need to remain practical for hands-on agent and MCP implementation. Prepare one example that traces a design decision through a working component and its trade-offs.
Sharing context without weakening controls
Likely preparation challenge: reusable RAG and context frameworks must make knowledge useful while supporting the posting’s governance and security-review duties. Prepare to discuss how you would assess access boundaries, retrieved-content quality and documentation choices.
Standardizing without blocking adoption
Likely preparation challenge: common agent skills, prompts and governance standards can improve reuse but may not fit every business use case. Prepare to describe what you would standardize, where you would allow variation and how you would explain that distinction to stakeholders.

Practice your answers

Technical practice

How would you design an MCP server architecture for an enterprise AI platform?

What to coverThe role explicitly includes designing and implementing MCP servers. Preparation suggestion: Explain how you would define server responsibilities and access boundaries, and how you would assess scalability and security trade-offs.

When would you use LangGraph rather than a simpler LangChain workflow for an AI agent?

What to coverLangChain, LangGraph and agentic AI are named requirements. Preparation suggestion: Compare straightforward workflows with agents that need explicit state and branching, and explain how you would troubleshoot a failed step.

How would you design and evaluate a RAG-based solution for enterprise knowledge?

What to coverThe posting calls for RAG-based solutions and enterprise knowledge architectures. Preparation suggestion: Cover content selection, retrieval and answer grounding, then describe how you would diagnose poor retrieval separately from poor answer generation.

What would an enterprise context-engineering framework need to standardize?

What to coverBuilding such a framework is a stated responsibility. Preparation suggestion: Discuss how an agent receives relevant instructions, retrieved material and permitted context, including trade-offs between relevance, context size and information exposure.

How would you make agent skills and prompt libraries reusable?

What to coverThe role includes developing reusable agent skills and prompt libraries, with prompt engineering and OpenAI or Claude APIs listed as mandatory skills. Preparation suggestion: Explain interface and versioning choices, and how you would test changes without assuming one prompt works equally well across tasks or APIs.

How would you divide implementation responsibilities across an Azure-based enterprise AI platform?

What to coverThe posting names Azure, Python and TypeScript or Node.js alongside hands-on platform architecture. Preparation suggestion: Sketch component boundaries and language choices, explaining their effects on maintainability, integration and scale without assuming a particular Azure service is required.

When would an ontology or knowledge graph improve an enterprise knowledge architecture?

What to coverOntology models are a stated design responsibility; knowledge graphs and semantic data architecture appear as preferred skills. Preparation suggestion: Contrast explicit relationships with simpler knowledge organization, and explain when modeling effort would or would not be justified.

How would you turn AI governance goals into architecture and security reviews?

What to coverThe role calls for AI governance standards and security review documentation. Preparation suggestion: Describe reviewable controls for access, data handling and agent behavior, including how you would verify that a proposed design follows them.

What would you record in an ADR for a consequential AI architecture choice?

What to coverWriting Architecture Decision Records and SDLC standards is an explicit duty. Preparation suggestion: Show how you would document options, trade-offs, the decision and its consequences, then connect the decision to implementation and review practices.

How would you investigate an agentic solution that works in a prototype but fails at enterprise scale?

What to coverThe role combines hands-on implementation with delivery of scalable AI solutions. Preparation suggestion: Separate possible failures in orchestration, context, retrieval and platform integration, then explain how evidence would guide a fix rather than a wholesale redesign.

Behavioural practice

Tell me about a time you aligned architecture, engineering and business teams on a technical design.

What to coverThe posting calls for collaboration across those teams. Preparation suggestion: Use a real situation, your actions and the result; focus on how you resolved differing priorities and made the decision understandable.

Describe a time you moved from an architecture decision into hands-on implementation.

What to coverThe role explicitly bridges strategic architecture and implementation. Preparation suggestion: Use a real situation, your actions and the result; show what implementation taught you and whether it changed the design.

Tell me about a time you introduced a shared technical standard that others were reluctant to adopt.

What to coverDriving AI platform adoption, governance and best practices is a stated responsibility. Preparation suggestion: Use a real situation, your actions and the result; explain how you addressed objections and observed whether adoption improved.

Describe a time a security or governance concern changed your proposed solution.

What to coverThe posting includes AI governance standards and security review documentation. Preparation suggestion: Use a real situation, your actions and the result; explain the trade-off and how the revised approach was reviewed.

Tell me about a time you documented a difficult architecture decision for teams that would maintain the system.

What to coverAuthoring ADRs and SDLC standards is part of the role. Preparation suggestion: Use a real situation, your actions and the result; focus on how the documentation clarified trade-offs or supported later implementation decisions.

Sharpen your knowledge

Topics to prepare

MCP and agent architecture
The posting requires MCP and agentic AI architecture. Suggested revision: sketch an MCP server and agent architecture, then explain how the components interact and where access boundaries belong.
Enterprise context engineering
The role calls for enterprise context-engineering frameworks. Suggested revision: outline how an agent would select, use and limit context from enterprise sources, and explain the trade-offs.
RAG and enterprise knowledge architecture
RAG solutions and enterprise knowledge architectures are responsibilities; knowledge graphs and ontology design are preferred skills. Suggested revision: compare a basic RAG design with an ontology-informed design for the same knowledge set.
Enterprise AI platforms on Azure
Enterprise AI architecture and Azure are mandatory. Suggested revision: diagram a scalable AI platform on Azure and explain its component boundaries and architecture decisions without assuming a particular Azure service is required.
AI governance and architecture standards
The posting calls for AI governance standards, ADRs, SDLC standards and security review documentation. Suggested revision: identify the decisions and review questions those artifacts should address for an agentic AI design.

Skills to revise

Implement an MCP server
MCP server implementation and Python are stated requirements. Suggested exercise: build a small Python MCP server that exposes a bounded enterprise-knowledge operation, then demonstrate how an agent would use it.
Build a reusable agent workflow
LangChain, LangGraph and reusable agent skills are named in the posting. Suggested exercise: implement a small LangGraph workflow using LangChain components, with a reusable agent skill and a clear path for handling an unsuccessful step.
Prototype RAG and prompt reuse
RAG, prompt engineering, prompt libraries and the OpenAI or Claude API are listed. Suggested exercise: create a small RAG flow using one of those APIs, save reusable prompt variants and compare responses against the retrieved material.
Implement a TypeScript or Node.js integration
The mandatory skills include Python and TypeScript/Node.js. Suggested exercise: write a small TypeScript or Node.js client for the agent workflow and show how it passes a request and handles a response.
Produce architecture and security artifacts
Authoring ADRs, SDLC standards and security review documentation is an explicit responsibility. Suggested exercise: write a one-page ADR for the prototype, plus a brief security review and SDLC checklist.
Plan delivery with the named platform tools
Azure, GitHub, Jira and Confluence are mandatory skills, and the role works across architecture, engineering and business teams. Suggested exercise: map the prototype to an Azure deployment plan and organize its code, delivery tasks and architecture notes using the named tools.

Your preparation checklist

Map the role’s requirements to evidence
Prepare a one-page evidence map for the posting’s architecture experience, degree and mandatory technologies. Note which items you can substantiate and which would need clarification; no candidate background was provided.
Prepare an enterprise AI project walkthrough
Select a real project, if available, and outline your architecture decisions, hands-on contribution, scale considerations and collaboration with engineering and business teams. Label any hypothetical example as hypothetical.
Sketch an MCP-based agent architecture
As interview practice, draw an enterprise-scale design showing MCP servers, agent workflows, LangChain or LangGraph orchestration, model API access and Azure deployment. Be ready to explain boundaries and trade-offs; this is a suggested exercise, not a stated interview task.
Work through context and retrieval design
Prepare a short design for context selection, reusable prompts or agent skills, and RAG over enterprise knowledge. Explain how you would evaluate answer quality and where an ontology or knowledge graph might help; those are preferred, not mandatory, skills.
Draft an architecture decision and security review
Create a sample ADR comparing two implementation choices, plus a brief security-review outline and SDLC considerations. These reflect the posting’s documentation and AI-governance responsibilities.
Organize implementation evidence
Identify specific examples you could discuss for Python, TypeScript or Node.js, an OpenAI or Claude API, GitHub, Jira and Confluence. For each, prepare what you built or documented and the technical decision it illustrates.
Prepare questions for the employer
Ask how enterprise architecture, engineering and business teams share decisions; what AI platform adoption would involve; and how the stated 12-month contract duration relates to the full-time listing and Gurgaon hybrid arrangement.

Availability last checked: 5 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…