Principal Architect - Agentic AI & AI-Native Software Engineering

Company Overview

Req ID: 395474

NTT DATA strives to hire exceptional, innovative, and passionate individuals who want to grow with us. If you want to be part of an inclusive, adaptable, and forward-thinking organization, apply now. We are currently seeking a professional to join our team in Mahwah, New Jersey (US-NJ), United States (US).

• AI-Native Engineering Operating Model:
  • Design the target operating model for small, outcome-oriented engineering pods that combine human product/domain expertise with specialized AI agents.
  • Define the optimal division of work between product/domain leaders, full-stack engineers, and AI agents across requirements, architecture, development, testing, security, release, operations, and documentation.
  • Create clear human-in-the-loop controls, approval points, escalation paths, and accountability for agent-generated work.
  • Partner with delivery leaders to transition fragmented staffing models into cohesive, accountable pods with end-to-end ownership of outcomes.
  • Define skill profiles and upskilling pathways that enable engineers to operate as AI-enabled full-stack product builders rather than narrow functional specialists.
• Agentic SDLC Architecture & Engineering:
  • Architect end-to-end agentic workflows spanning discovery and requirements, solution design, code generation, code review, test generation and execution, security checks, CI/CD, deployment, observability, incident analysis, and technical documentation.
  • Design and deploy autonomous and semi-autonomous agents with planning, reasoning, memory, context management, tool/function calling, and multi-agent collaboration capabilities.
  • Implement agent orchestration patterns that allow multiple specialized agents to collaborate reliably while maintaining deterministic controls for high-risk actions.
  • Integrate agents with engineering platforms and enterprise systems such as GitHub, Azure DevOps, JFrog, cloud services, observability platforms, service management tools, databases, APIs, and internal knowledge sources.
  • Ensure agentic workflows are designed for production use, including resilience, fallback behavior, auditability, traceability, and secure access to enterprise systems.
• Reusable Agentic Engineering Blueprints:
  • Identify successful AI-enabled and agentic engineering patterns already demonstrated within delivery teams and convert them into reusable reference architectures, playbooks, templates, and accelerators.
  • Create standardized blueprints for common engineering use cases including requirements decomposition, story creation, coding, self-review, test automation, defect analysis, security analysis, log analysis, release automation, and operational support.
  • Design the enablement approach required to replicate proven patterns across multiple products, portfolios, and engineering pods without creating bespoke implementations for every team.
  • Establish a pattern library and reusable agent catalog with clear ownership, versioning, guardrails, and adoption guidance.
• Engineering Productivity & Business Outcome Measurement:
  • Establish productivity baselines before agentic transformation and define measurable target improvements for each pod.
  • Quantify changes in throughput, cycle time, lead time, time-to-production, release frequency, quality, escaped defects, engineering effort, cost, and roadmap delivery.
  • Design controlled experiments that compare traditional, AI-assisted, and agentic delivery models and isolate the impact of AI-enabled ways of working.
  • Translate engineering productivity improvements into executive-level business outcomes, such as accelerated roadmap milestones, increased feature throughput, reduced operating cost, improved service quality, and faster realization of business value.
  • Build executive dashboards that show pod-level and portfolio-level AI productivity, ROI, adoption, quality, and outcome metrics.
• Context Engineering, Knowledge & Legacy Modernization:
  • Design enterprise context architectures that give agents reliable access to product knowledge, business rules, architecture standards, source code, technical documentation, and operational data.
  • Architect RAG, knowledge graph, semantic search, and persistent context-store patterns that improve agent accuracy and reduce repeated context loading.
  • Apply context engineering to modernization and reimagination efforts while preserving the role of produ…