Job ID: 111076

Gurugram

Do you want to do work that matters, alongside supportive leaders who will help you grow faster than you ever thought possible? Are you a creative problem-solver who is energized by challenges? You’ve come to the right place.

YOUR IMPACT

In this role, you will define and lead the technical vision for intelligent agent-based systems, ensuring they are designed, orchestrated, and scaled to meet the complex needs of the enterprise.

You will architect multi-agent systems that integrate large language models (LLMs), APIs, tools, and memory systems to enable autonomous and semi-autonomous workflows. By establishing robust frameworks, operational guardrails, and safety mechanisms, you will ensure these systems are secure, reliable, and compliant with governance and ethical standards.

Your focus will be on building scalable platforms that incorporate advanced capabilities such as retrieval-augmented generation (RAG), structured memory systems, and seamless integration with enterprise tools and data sources. You will collaborate closely with engineering and product teams, providing technical leadership, defining best practices, and guiding the development and deployment of innovative AI-driven solutions.

In this role, you will play a critical part in shaping the future of AI systems within the organization, driving transformative impact by enabling intelligent, scalable, and secure agent-based technologies that unlock new possibilities for business-critical workflows.

Agentic Data Mesh Architecture

- Design and lead an Agentic Data Mesh architecture in which intelligence moves across autonomous product domains while data remains within domain-owned systems.

- Architect a Master Agent that understands user intent, maintains conversation memory, discovers domain capabilities, plans execution, orchestrates MCP calls, and synthesizes cross-domain responses.

- Define the architecture and standards for domain agents and MCP servers, with each domain retaining ownership of its data, semantic layer, business methodology, data retrieval, and authorization.

- Establish capability and client registries that enable dynamic discovery of MCP servers, capabilities, entities, inputs, outputs, and canonical client identities.

- Design semantic layers that translate enterprise business concepts into domain-specific queries and calculations, enabling reliable cross-domain reasoning without centralized data consolidation.

- Build for enterprise-grade security and resilience through dual authorization, guardrails, retries, graceful degradation, partial-result handling,



failure logging, and support workflows.

- Create a composable platform model in which recent product domains can join the mesh by exposing standardized MCP capabilities without requiring central platform re-architecture.

- Partner with product and engineering leaders to establish architectural principles, integration standards, observability, governance, and operating practices for a scalable multi-agent ecosystem.

What You Will Architect

- Master Agent / orchestration layer

- Domain agents and MCP interfaces

- Per-domain semantic and methodology layers

- Client and capability registries

- Conversation memory and contextual multi-turn orchestration

- Cross-domain planning, tool calling, synthesis, and failure handling

- Security, authorization, governance, and AI guardrails

- Scalable event-driven and distributed integration patterns

YOUR GROWTH

Driving lasting impact and building long-term capabilities with our clients is not easy work. You are the kind of person who thrives in a high performance/high reward culture - doing hard things, picking yourself up when you stumble, and having the resilience to try another way forward.

In return for your drive, determination, and curiosity, we'll provide the resources, mentorship, and opportunities you need to become a stronger leader faster than you ever thought possible. Your colleagues—at all levels—will invest deeply in your development, just as much as they invest in delivering exceptional results for clients. Every day, you'll receive apprenticeship, coaching, and exposure that will accelerate your growth in ways you won’t find anywhere else.

When you join us, you will have:

- Continuous learning: Our learning and apprenticeship culture, backed by structured programs, is all about helping you grow while creating an environment where feedback is clear, actionable, and focused on your development. The real magic happens when you take the input from others to heart and embrace the fast-paced learning experience, owning your journey.
- A voice that matters: From day one, we value your ideas and contributions. You’ll make a tangible impact by offering innovative ideas and practical solutions,



all while upholding our unwavering commitment to ethics and integrity. We not only encourage diverse perspectives, but they are critical in driving us toward the best possible outcomes.
- Global community: With colleagues across 65+ countries and over 100 different nationalities, our firm’s diversity fuels creativity and helps us come up with the best solutions for our clients. Plus, you’ll have the opportunity to learn from exceptional colleagues with diverse backgrounds and experiences.
- World-class benefits: On top of a competitive salary (based on your location, experience, and skills), we provide a comprehensive benefits package to enable holistic well-being for you and your family.

YOUR QUALIFICATIONS AND SKILLS

Bachelor’s degree in Computer Science, Engineering, or a related field (Master’s preferred)

7+ years of software/platform architecture experience, including 1+ year with LLMs or AI platforms

Experience building AI agents, autonomous workflows, prompt engineering, tool calling, and function execution

Hands-on experience with vector databases, embeddings, RAG, distributed systems, and APIs

Experience with Kubernetes, cloud platforms, and workflow engines (Temporal, Airflow, Step Functions)

Familiarity with LangGraph, AutoGen, CrewAI, or Semantic Kernel

Experience implementing AI governance, security, compliance, and guardrails

Knowledge of AI observability, prompt tracing, cost tracking, and failure analysis

Expertise in scalable, secure, event-driven architectures and workflow orchestration

Strong collaboration and communication skills across engineering, product, legal, and risk teams

Ability to balance domain autonomy with enterprise-wide consistency, governance, security, observability, and reliability.

Experience designing semantic layers, metadata/capability registries, identity models, authorization boundaries, and enterprise data-access patterns.

Strong understanding of multi-agent orchestration, agent-to-agent/tool interfaces, MCP-style capability discovery, and separation of orchestration from domain business logic.

Experience architecting distributed, domain-oriented data or AI platforms, preferably using data mesh, federated architecture, or similar decentralized patterns

Please review the additional requirements regarding essential job functions of McKinsey colleagues.

Our unwavering commitment to integrity drives everything we do, guiding us to always act in the best interests of our clients, our people, and the communities we serve.

📌 Principal AI Architect (Gurugram)
🏢 McKinsey
📍 Gurugram