Union Home Mortgage is a $5B+ annual originator building internal ML and data infrastructure to reduce manual loan processing work. The tech stack reveals a data-science-forward operation: Python, R, TensorFlow, PyTorch, MLflow, and Kubeflow sit alongside mortgage-specific tooling (Encompass). Active projects on document classification, data extraction, and compliance-ready AI solutions, paired with pain points around loan manufacturing costs and processing delays, suggest the company is moving from manual workflows toward automated underwriting and quality control.
Union Home Mortgage originates residential mortgages across 44 states and D.C., serving hundreds of thousands of borrowers since 1970. The company operates at scale—roughly $5 billion in annual lending volume—with a headquarters in Strongsville, Ohio and a 1,000–5,000-person workforce. The current hiring mix is heavily skewed toward loan officers and sales roles (41 of 53 active postings), with a small engineering team (6 roles) and operational support. Product focus includes both customer-facing mortgage origination and internal infrastructure for compliance, risk, and cost reduction.
Python, R, C#, .NET, TensorFlow, PyTorch, scikit-learn, Apache Airflow, Docker, AWS, Azure, MLflow, Kubeflow, SageMaker, and Encompass mortgage platform. Also Microsoft 365 suite, Active Directory, and Azure DevOps.
ML-driven loan automation: document classification, data extraction models, end-to-end data validation pipelines, and compliance-ready AI solutions. Also branch expansion and loan officer training infrastructure.
Other companies in the same industry, closest in size