Principal AI Engineer

Who is Stanley Martin Homes? At Stanley Martin Homes, we believe your work should have a purpose. With us, it truly does. Our success starts with our people, and we are proud to foster a culture where every team member is valued and supported. At Stanley Martin Homes, you will work alongside passionate, knowledgeable professionals who are committed to doing the right thing, delivering exceptional homebuyer experiences, and putting homebuyers first. Stanley Martin Homes is one of the largest 25 homebuilders in the United States and it has been consistently one of the fastest growing. We are proud of the people-first culture that makes it possible. If you are ready to build a meaningful career and help families find the place they will call home, we would love to connect with you. Join our team and build a career that you will be proud of. Explore Opportunities Today The Principal AI Engineer is a role where the domain is the entire business. The Principal AI Engineer will work directly with teams across Finance, Team and Culture, Operations, Sales, Purchasing, and Land to understand how work gets done, identify where agentic AI can remove manual effort and improve decisions, and build the production systems that make it real. The Principal AI Engineer will go deep into enterprise agentic AI toolchains and coding agents (e.g., Claude Code), including configuration, subagents, tool and connector integration, skill authoring, and evaluation. The Principal AI Engineer will establish the platform, standards, and reference patterns that every future agent builder at Stanley Martin will follow. Responsibilities and Duties Agentic AI Platform & Enablement • Stand up and own Stanley Martin’s agentic AI platform: enterprise agentic tooling and coding-agent configuration, API access, MCP (Model Context Protocol) servers and secure connectors to core business systems (ERP, cloud data platform, HR, finance, and collaboration systems), and the secure development environment. • Operationalize the company’s secure AI adoption framework as the baseline for all agent development. • Build out the enterprise context layer in partnership with Business Technology: turn institutional knowledge, process documentation, business rules, and system metadata into governed, machine-readable context (knowledge bases, semantic models, and MCP resources) that agents rely on to reason accurately about how Stanley Martin works. • Partner with EDAP and business stakeholders to operationalize semantic models, business definitions, and enterprise context required for trusted AI solutions. • Build and maintain a version-controlled repository of reusable skills, agent templates, and reference architectures. • Define the standard patterns, starter kits, and guardrails that let EDAP, DEV, and other teams and users build agentic solutions safely on the platform. Agent Design & Delivery • Design, build, and ship production agentic systems end-to-end, from discovery and scoping through architecture, implementation, evaluation, and deployment. • Deliver the team’s first production agents in partnership with executive sponsors across business functions. • Implement core agentic patterns: tool and function calling, planner-executor flows, structured output, multi-step reasoning, retrieval-augmented generation, and human-in-the-loop controls. • Build hybrid solutions where agents reason and decide, then call deterministic automations to execute, partnering with Business Technology on the automation layer. • Operate across a heterogeneous environment that includes cloud infrastructure, an enterprise data platform, productivity and collaboration systems, Microsoft Dynamics ERP, and 3rd party SaaS platforms (e.g., Salesforce). Evaluation, Review & Governance • Build the evaluation harness every agent must pass before production: test cases, accuracy thresholds, and quality regression testing. • Establish and run the formal agent review process, covering intake, design and data-classification review, security review, human-in-the-loop tier assignment, and promotion gates from development to production. • Implement observability for deployed agents, including audit logging of agent actions, telemetry on usage, errors, latency, and drift, and scheduled post-deployment reviews. • Implement cost and value controls, including usage monitoring and model routing. • Ensure every solution aligns with Stanley Martin’s AI policies, data classification standards, least-privilege access design, and responsible AI guardrails. Shareability, Catalog & Federation • Build and curate the central agent and skill catalog, so agentic capabilities are discoverable and reusable across the company rather than rebuilt. • Publish documentation standards and contribute reusable connectors and skill packages. • Support EDAP (data-platform AI capabilities), DEV (workflow automation and system integrations), and other teams so they build on the platform under shared standards. • Lead internal showcases and demos that drive adoption across divisions and functions. Cross-Functional Partnership & Technical Leadership • Partner directly with business stakeholders to map real workflows, capture institutional knowledge, and identify high-leverage agentic AI opportunities. • Help triage incoming ideas between agentic AI and conventional automation, routing deterministic work to the right lane. • Influence technical direction by setting architecture patterns, reviewing designs, and raising engineering standards. • Mentor engineers and builders across the company on agentic patterns and best practices as adoption grows. • Stay current on advancements in agent frameworks, evaluation approaches, and responsible AI governance, and translate them into Stanley Martin’s roadmap. • Complete all other duties as assigned by manager. • Represent the company professionally in all internal and external interactions and communications. • Adhere to safety standards and help promote a safe working environment. • Adhere to and promote the Mission, Vision, and Values of Stanley Martin Homes Position Standards • Experience designing APIs, services, and data pipelines. You think in systems, not scripts. • Ability to operate with high autonomy. You scope your own work, make architectural decisions, and own outcomes. • Excellent communication skills, with the ability to explain technical concepts and trade-offs to non-technical executives and business partners. • Willingness to work across the full range of business systems, including the unglamorous ones where the high-impact problems live. Position Requirements • 8+ years of professional engineering experience, with a track record of designing, shipping, and operating production systems. • Bachelor's degree in computer science or a related field, or equivalent practical experience. • Hands-on experience building with LLM APIs, agent frameworks, or AI-powered applications that ran in production or were used by real people, including agentic patterns such as tool use, orchestration, and human-in-the-loop design. • Depth with modern agentic AI toolchains and coding agents including Anthropic Claude, Snowflake, Omni, and MCP-based architectures, or a demonstrated ability to develop that depth quickly, including connector and tool integration, skill and subagent design, and prompt engineering. • Cloud fluency, with Microsoft Azure strongly preferred; experience with enterprise cloud data platforms (e.g., Snowflake or comparable). • Evaluation and testing discipline: you can define what good means for an agent and prove it before production. • Security- and governance-minded: comfortable with data classification, least-privilege design, auditability, and human-in-the-loop controls. • Strong proficiency in Python; comfort with additional languages (TypeScript/JavaScript, SQL) is a plus. • Willingness to work in a hybrid environment (3 days per week onsite at Reston, VA headquarters). Preferred Qualifications • Exper