Lead Machine Learning Engineer

Lennar•Boston, MA

2d•$152,600 - $190,700•Remote

About The Position

Lennar is seeking a Machine Learning Engineer to own and evolve the infrastructure and surface mechanisms that take our data science and ML models from notebook to production. This is a key role on the Applied AI & Data Science team, sitting at the intersection of software engineering, ML platform, and applied data science. The ideal candidate is a software engineer with deep MLOps expertise. They know how to design model serving for both batch and real-time inference, build durable model registries and versioning practices, and stand up retraining pipelines that data scientists actually use. They are hands-on with AWS SageMaker (including SageMaker Unified Studio), MLflow, Weights & Biases, and the surrounding tooling that makes ML systems reliable in production. You’ll partner closely with data scientists, AI engineers, and platform teams—building and setting the foundation that lets ML models ship faster, retrain on schedule, and operate with the same engineering rigor as any other production service across 40+ divisions of one of the nation’s largest homebuilders.

Requirements

Bachelor’s degree or higher in Computer Science, Engineering, or a related technical field.
7+ years of software engineering experience, including meaningful production ownership of services or platforms in a cloud environment.
5+ years of hands-on MLOps or ML platform experience—deploying, monitoring, and retraining production models at scale.
Strong hands-on experience with AWS SageMaker (Unified Studio strongly preferred), including model training jobs, endpoints, batch transform, and pipelines.
Deep experience with experiment tracking, model registries, and retraining workflows using MLflow, Weights & Biases, or comparable tooling.
Strong Python skills with a track record of writing modular, well-tested, production-ready code; experience with infrastructure-as-code (Terraform preferred).
Solid understanding of both batch and real-time inference patterns, including the tradeoffs between latency, throughput, cost, and operational complexity.
Proven ability to partner with data scientists—understanding their workflow, lowering friction, and translating modeling needs into reliable platform capabilities.
Comfortable operating with autonomy in ambiguous environments—scoping work, setting realistic timelines, and raising blockers proactively without waiting to be asked.

Nice To Haves

Experience with feature stores, model gateways, GPU workloads, distributed training, model drift monitoring tools, or supporting both classical ML and LLM-based models on the same platform.

Responsibilities

Design, build, and set the ML platform surface used by our data science team—covering model packaging, deployment, batch and real-time inference, and observability.
Establish and evangelize ML platform standards, patterns, and reusable components—raising the engineering bar for how ML models are built, deployed, and operated across the organization.
Mentor data scientists and engineers on production ML practices, code review their platform-adjacent work, and serve as the technical authority on MLOps decisions.
Own model serving infrastructure on AWS SageMaker (including SageMaker Unified Studio)—building patterns for batch inference jobs, real-time endpoints, and serverless inference depending on workload requirements.
Build and maintain the model registry, version control, and promotion workflows that move models cleanly from development to staging to production with full lineage and auditability.
Stand up and operate retraining pipelines using MLflow, Weights & Biases, and orchestration tools—automating retraining triggers, experiment tracking, model evaluation, and approval gates.
Build monitoring and alerting for production models including drift detection, performance degradation, data quality issues, and latency or cost anomalies.
Write clean, modular Python and infrastructure-as-code (Terraform) for ML platform components, applying software engineering best practices including testing, versioning, and code review.
Partner closely with data scientists to make their workflow faster and more reliable—reducing time-to-production for new models and increasing confidence in models already in production.
Collaborate with Data / Platform Engineering and AI Engineering counterparts to ensure feature pipelines, model artifacts, and inference services are integrated cleanly with the broader data and AI platform.