About the Role

HDJ & Associates is representing a client seeking a Senior Principal Architect to lead an Agentic AI initiative across customer-facing digital platforms. This role involves designing scalable, governed systems for loyalty, payments, and commerce workflows while integrating AI agents with enterprise systems.

Skills

Agentic AI LLM LangGraph LangChain Semantic Kernel CrewAI AutoGen Python RAG MCP LLMOps System Architecture Multi-agent Orchestration Retail Systems Integration Prompt Engineering Observability

Full job details

Our client is looking for a Senior Principal Architect Agentic AI & Digital Platforms to serve as the onshore technical anchor for an Agentic AI initiative across customer-facing digital platforms. The ideal candidate will translate business intent into scalable, governed, production-ready systems and hold the architectural line across delivery teams. The right person codes, reviews, and debates trade-offs in the same week they brief a VP.

Key Responsibilities:

Agentic Architecture - Design & govern end-to-end agentic architecture for apps, loyalty, payments, and commerce workflows.
Multi-Agent Orchestration - Orchestrator/subagent hierarchies, tool calling, stateful graph workflows, agent memory layers.
MCP Standards - Establish Model Context Protocol standards for governed tool and enterprise service exposure.
Runtime Guardrails - Spend controls, fraud hooks, human-in-the-loop checkpoints, and full auditability trails.
Retail Systems Integration - Lead AI agent integration with catalog, inventory, pricing, promotions, loyalty, and checkout.
LLMOps Standards - Prompt versioning, RAG governance, evaluation harnesses, observability (latency, cost, drift, safety).

Requirements

Must-have (non-negotiable):

12–18 yrs enterprise/platform architecture
3–4+ yrs AI/ML or GenAI in production
Shipped LLM-based agent systems (not POC)
Hands-on: LangGraph / LangChain / Semantic Kernel / CrewAI / AutoGen
ReAct, multi-agent hand offs, RAG, MCP knowledge
Architected against hallucination, tool misuse, cost runaway
Python proficiency + reference implementations
Translates tech → business value for VPs

Preferred:

Direct retail / commerce domain exposure
LLMOps depth (LangSmith / Langfuse)
Onshore/offshore delivery leadership
Knowledge graphs as grounding layers