Senior Principal Architect - Agentic AI & Digital Platforms
(FDE – Forward Deployment Engineer)
Dallas, TX (Preferred) | Onsite| 12–18 Years Experience
Role Overview
A hands-on techno-functional architecture role - not research, not purely strategic. The Senior Principal Architect serves as the onshore technical anchor for an Agentic AI initiative across customer-facing digital platforms: translating business intent into scalable, governed, production-ready systems and holding 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
- Design and govern end-to-end agentic architecture for digital experiences spanning apps, loyalty, payments, and commerce workflows.
- Architect multi-agent orchestration patterns - orchestrator/subagent hierarchies, tool/function calling, stateful graph-based workflows, and agent memory (short-term, session, long-term).
- Establish MCP (Model Context Protocol) standards for governed tool and enterprise service exposure.
- Define runtime guardrails: spend controls, fraud detection hooks, human-in-the-loop checkpoints, and auditability trails.
- Lead AI agent integration with retail systems - catalog, inventory, pricing, promotions, loyalty, and checkout.
- Set LLMOps standards: prompt versioning, RAG pipeline governance, evaluation harnesses, and observability (latency, token cost, drift, safety).
- Serve as the primary technical interface between business leaders, delivery teams, and platform partners.
Required Qualifications:
Foundational
- 12–18 years in enterprise/platform architecture; 3–4+ years delivering AI/ML or GenAI systems in production.
- Strong command of API-first design, distributed systems, event-driven architecture, and real-time data.Expertise in building and scaling applications.
Agentic AI - Core Requirement
- Production experience with LLM-based agent systems - shipped, not just prototyped.
- Hands-on with agent frameworks: LangGraph, LangChain, Semantic Kernel, CrewAI, or AutoGen - and judgment on agentic architectural patterns
- Working knowledge of: ReAct/plan-execute patterns, multi-agent handoffs, RAG pipeline design (chunking, hybrid retrieval, reranking), context/prompt engineering, and MCP.
- Familiarity with agent failure modes - hallucination, tool misuse, cost runaway, infinite loops - and how to architect against them.
- LLMOps exposure: evaluation frameworks, observability tooling (LangSmith, Langfuse, or equivalent), prompt/model lifecycle management.
Retail / Consumer Domain
- Meaningful exposure to retail commerce systems: catalog, pricing, promotions, loyalty, checkout, or payments. Direct retail experience strongly preferred; adjacent commerce background considered.
Data & Integration
- Comfortable with event-driven architectures (Kafka or equivalent) and enterprise data quality realities.
- Familiarity with semantic models or knowledge graphs as agent grounding layers is a plus.
Leadership Profile
- Translates agentic AI concepts into business value language credibly at the executive level.
- Sets architectural boundaries, makes trade-off calls, and defends them under pressure.
- Experienced leading across onshore/offshore delivery models.
- Hands-on enough to write reference implementations and lead code reviews - Python proficiency expected.
- Comfortable operating in a fast-moving technology space with evolving standards.