Director of Product – AI Underwriting

Arive•San Ramon, CA

5h•Remote

About The Position

ARIVE AI brings agents into the core of the platform - built in, not bolted on. Its underwriting agent runs guideline-grounded rules across live loan files to produce Pre-Underwrite outputs: findings, checklists, and conditions that surface eligibility issues and documentation gaps. The agents also try to autonomously address any issues before a file ever reaches the lender underwriting. The result is cleaner files into every lender’s process - and it only works if the rules behind it are right. The Director of Product - AI Underwriting owns the correctness of ARIVE AI’s Underwriting Rules and Pre-Underwrite outputs. This is an underwriting-expert-first role for someone who also understands the impact of AI: you bring deep, hands-on underwriting analysis and working AI fluency — reviewing prompts, shaping evaluation criteria, and reasoning about agent behavior — while our engineering team owns the low-level LLM development and evaluation infrastructure. You will verify that every rule ARIVE AI generates and applies is faithful to the agency guides and to each lender’s specific guidelines and overlays, working directly with lenders and agencies to review rules — and you will stand behind what the Pre-Underwrite says. This is a remote position with core Pacific Time working hours, working closely with the ARIVE platform engineering team and reporting to executive leadership and to the Head of Product once that role is hired. You will start as a hands-on individual contributor, with the potential to grow a team as the domain scales.

Requirements

Deep, demonstrated experience in detailed underwriting analysis — full-file review spanning the income, asset, credit, and collateral disciplines above — at senior scope.
10+ years in mortgage underwriting and/or agency credit policy.
Mastery of agency guidelines — Fannie Mae Selling Guide, Freddie Mac Seller/Servicer Guide, and FHA Handbook 4000.1 — with VA and USDA experience valued.
Working knowledge of AI/LLM platforms and tools, and a genuine passion for technology.
Should be able to review prompts, reason about agent behavior, and contribute to evaluations — plus familiarity with modern underwriting technology (AUS, income/asset verification tools, AI rules/analysis engines).
Willingness to travel approximately 20% — to our San Ramon office and to lender and agency partners.

Nice To Haves

A Bachelors or Masters degree is a strong plus.
Non-QM underwriting experience strongly preferred.
Wholesale/TPO or correspondent lending background strongly preferred, including collaboration with lender underwriters, account executives, setup teams, etc.
Bay Area–local candidates able to work regularly from the office are a strong plus.

Responsibilities

Review, test, and verify ARIVE AI Underwriting Rules against the agency guides and against each investor’s guidelines and overlays. Approve rules before release and re-verify them as guidelines change.
Verify the agents’ autonomous actions with the same rigor - every automated issue resolution must be correct and guideline-compliant.
Work directly with lender credit and underwriting teams, and with the agencies as applicable, to confirm rule accuracy and correctness — run validation reviews, leverage findings, reconcile interpretation differences, and maintain lender-specific rule sets as the source of truth.
Be responsible for Pre-Underwrite outputs end to end: audit findings, checklists, conditions, and the agents’ autonomous issue resolutions for accuracy on real files; triage errors and drive root-cause fixes; own the accuracy bar for what “correct” means.
Direct the rules’ treatment of income, asset, credit, and collateral underwriting across agency and Non-QM programs - the full expertise map below.
Work daily with platform engineers — who own low-level LLM development and evaluation infrastructure - to review and refine agent prompts, build evaluation datasets from the guides, design human-in-the-loop review and escalation, and validate agent output on messy, real-world files.
Monitor agency and lender guideline updates and drive timely, verified rule changes so Pre-Underwrite outputs never drift from current policy.
Shape the AI Underwriting roadmap and partner closely with Product, Engineering, Sales, Onboarding, Support, Compliance, and Executive Leadership to successfully launch new capabilities, integrations, and strategic initiatives
Define and report the domain’s quality KPIs — rule accuracy, false findings, condition quality, and lender acceptance of Pre-Underwrite outputs.