Principal AI Engineer

Lennar•Miami, FL

About The Position

The Principal AI Engineer is the top technical individual contributor in Lennar Technology Group's AI Engineering function. The role sits within the Data & AI organization, in the Applied AI & Data Science team, and is responsible for the architecture, scalability, and technical integrity of the foundational systems that power Lennar's AI future. This person sets the technical direction for the platforms that turn AI from experimentation into a durable enterprise capability — Model Context Protocol (MCP) servers that connect agents to Lennar's business systems, the AWS-native runtime that hosts them, the Skills Marketplace and Agent Templates that allow product teams to stand up new agents quickly, the BigLen platform, and the evaluation and observability infrastructure that keeps every model and agent measurable, governed, and safe. The Principal AI Engineer operates as a force multiplier — defining standards, leading the most complex builds, and elevating the engineering capability of the entire team.

Requirements

Required Bachelor's degree in Computer Science, Engineering, Mathematics, or related technical field; advanced degree a plus.
12+ years of progressive software engineering experience, with a track record of designing and operating distributed systems in production at enterprise scale.
5+ years building, deploying, and operating AI/ML systems in production — agentic systems, LLM applications, RAG pipelines, or large-scale ML platforms.
Deep expertise in AWS, with hands-on experience across Bedrock or Bedrock AgentCore, Lambda, ECS/EKS, IAM, API Gateway, and VPC networking.
Strong programming proficiency in Python and TypeScript/Node.js; comfort with infrastructure-as-code (Terraform or CDK).
Working knowledge of the Model Context Protocol (MCP) ecosystem — server and client patterns, tool exposure, transport, and auth — or comparable experience building agent-tool integration frameworks.
Demonstrated experience building developer platforms, internal SDKs, or reusable framework infrastructure consumed by multiple downstream teams.
Experience designing and operating evaluation and observability frameworks for AI systems (LLM evals, traces, prompt and response capture, cost telemetry).
Strong grounding in software security, identity, and data protection — secrets management, least-privilege access, PII handling, encryption at rest and in transit.
Track record of technical leadership without people management — setting architecture, leading design reviews, mentoring senior engineers, and influencing cross-team decisions.
Excellent written and verbal communication; ability to translate complex technical decisions into clear language for executive audiences.

Nice To Haves

Experience with agent orchestration frameworks (LangGraph, AutoGen, CrewAI, or equivalent) and with production retrieval-augmented generation systems.
Familiarity with model gateways, prompt management, and feature/semantic store architectures.
Prior experience operating in an enterprise environment with mature governance, audit, and compliance requirements.
Exposure to Snowflake, dbt, and modern data platform patterns.
Experience with Coralogix or comparable observability platforms (Datadog, New Relic, Dynatrace) for distributed tracing and metrics.
Open-source contributions to AI infrastructure, agent frameworks, or MCP-related projects.

Responsibilities

AI Platform Architecture: Own end-to-end technical architecture for Lennar's enterprise AI platform — MCP server framework, agent runtime, tool and skill registries, model gateway, and orchestration layers — ensuring scalability, reliability, and cost efficiency at enterprise scale.
MCP Server Engineering: Design, build, and operate MCP servers that expose Lennar business systems (data platforms, productivity tools, line-of-business applications) to agentic clients with consistent auth, rate limiting, observability, and governance patterns.
AWS Infrastructure: Architect and operate AWS-native infrastructure (Bedrock, AgentCore, Lambda, ECS/EKS, API Gateway, IAM, VPC, CloudWatch) with a strong emphasis on infrastructure-as-code, secure-by-default patterns, multi-environment promotion, and cost transparency.
Skills Marketplace & Agent Templates: Lead the engineering of the internal Skills Marketplace and Agent/MCP Template platform — a versioned, discoverable, governed catalog of reusable agent capabilities and patterns that accelerates time-to-production for new AI products across LTG.
BigLen Platform: Own the foundational engineering layer of BigLen — the MCP routing fabric, contextual intelligence services, and decision-guidance pipelines — partnering with AI Products and AI Architecture to evolve the platform across its Phase I–III roadmap.
AI Evaluation Framework: Establish the enterprise evaluation framework for agents, tools, and models — automated regression tests, accuracy and quality benchmarks, safety and policy checks, and gating workflows that govern promotion from development to production.
AI Observability: Partner with the AIOps & Observability team to instrument every agent, tool, model, and skill with traces, metrics, cost telemetry, prompt and response capture, and policy signals; help operationalize a single pane of glass for AI system health.
Technical Standards & Patterns: Define the internal SDKs, reference implementations, code review standards, and architectural patterns that the AI Engineering team builds against. Set the bar for code quality, testing, and engineering rigor.
Security, Privacy, and Responsible AI: Embed AI Trust & Governance and Data Trust requirements into the platform by default — PII handling, data residency, access control, audit logging, model and prompt safety — in close partnership with the AI Governance and Data Protection teams.
Technical Leadership: Lead the highest-impact and most complex engineering work across Agentic Engineering, AI Products, and AI Architecture. Mentor senior and lead engineers, raise the technical level of the team, and represent the AI Engineering function in cross-functional and executive forums.
Build vs. Buy Judgment: Evaluate emerging models, frameworks, vendors, and open-source projects; lead build-vs-buy decisions with clear reasoning on cost, risk, lock-in, and strategic fit.