Staff Machine Learning Engineer

Doma Technology LLC offers solutions that simplify real estate closings for lenders, real estate professionals, title agents, and homeowners. The Staff ML Engineer will build and deploy scalable, production-grade machine learning systems for risk assessment, underwriting, and operational efficiency, while collaborating with technical and business stakeholders in a regulated domain.

Responsibilities:

• Design, build, and maintain scalable ML systems across the full model lifecycle—from data ingestion and training to deployment, monitoring, and retraining
• Serve as a go-to expert in title risk and underwriting for our team and the company. This includes exploring ways to reduce model losses, improve the fidelity of our projected losses, and identify new risks in our domain
• Collaborate with data engineers, product teams, and domain experts to translate business goals into ML solutions that perform in real-world production environments
• Conduct ad-hoc analyses and experimentation to inform modeling decisions and stakeholder strategy

Requirements:

• 6+ years developing and deploying models in a cloud-based production environment following MLOps and engineering best practices, with proficiency in Python, SQL, Git, and Docker
• Depth in supervised learning on tabular data — feature engineering, class imbalance, label definition, and threshold selection with the business tradeoffs behind it
• Experience monitoring and evaluating models in production over time — tracking degradation, drift, and error rates, and turning that into a defensible account of what changed and why
• Strong applied statistics. You can decompose a metric change into its drivers and defend the decomposition
• You're the reviewer, not the reviewed. You look for the reason a number might be wrong before a client's reviewer does, and you have a track record of catching problems in other people's analyses and raising the bar around you
• Fluent with current AI tooling in your own work**.** You already heavily use LLM-based coding and analysis tools, you have opinions about where they help and where they fail, and you keep up with what's changing. We're not looking to persuade anyone that these tools are worth using
• Track record of owning complex, ambiguous projects end to end and working directly with product and senior leadership — including pushing back on requirements when they're wrong