Mortgage AI Evaluation Expert (English)

You will use professional mortgage and home lending knowledge to review large language model responses. Your work will help assess whether AI answers are accurate, complete, compliant, and useful.

You will work with AI researchers and other lending professionals to create evaluation standards and benchmarks. Prior AI training or model evaluation experience is not required.

• Evaluate AI responses about residential underwriting, Conventional, FHA, VA, and USDA loan programs.
• Review responses covering refinancing, servicing, and the mortgage lifecycle from application through closing.
• Create rubrics for DTI and LTV calculations, borrower income and asset verification, appraisal review, escrow analysis, and TRID and RESPA disclosures.
• Identify errors, missing information, and regulatory issues in model responses.
• Help shape evaluation strategies, training methods, and benchmarks for mortgage-related AI systems.

What It Pays And Takes

This is a remote freelance contract with a flexible part-time commitment. The listing is marked entry level, and the role does not require previous AI training experience.

• $100 per hour.
• 20 or more hours per week.
• Remote work open worldwide.
• Professional fluency in English.
• At least two years of mortgage or home lending experience as an underwriter, loan officer, loan processor, closer, servicing specialist, secondary marketing analyst, or in a comparable role.
• Strong knowledge of Fannie Mae and Freddie Mac guidelines, mortgage underwriting standards, regulatory requirements, and the full loan lifecycle.
• Clear written communication, strong critical thinking, and the ability to explain lending judgments consistently.
• NMLS licensure, DE or LAPP underwriting authority, Certified Mortgage Banker certification, or an MBA in Finance is helpful but optional.

How It Works

Apply on OpenTrain with your resume, then complete the application on the hiring site.

About AI Training Work

AI training work is the human side of building artificial intelligence. People review and rate model responses so AI systems can become more accurate, useful, and reliable, and experienced professionals contribute the subject knowledge needed for specialized evaluations.