Outlier is an AI training platform owned and operated by Scale AI that connects a global community of subject‑matter experts with leading AI companies to improve the intelligence, reliability, and safety of advanced models. Scale AI provides the underlying data infrastructure, annotation services, and evaluation tools that power Outlier, enabling organizations to build and refine AI systems across domains such as autonomous vehicles, robotics, and language technologies.
Key responsibilities
- Receive and review initial construction packages.
- Ensure required documentation is obtained for each draw request.
- Work directly with clients to resolve complex documentation or funding issues.
- For each draw, collect and assess inspection/progress photos; reconcile percent complete to funded-to-date and remaining to complete.
- Calculate eligible advances for each construction draw.
- Monitor and escalate documentation gaps and exceptions to the PM and RM, as applicable.
- Partner with associates (PMs, RMs, Operations) and clients’ funding teams to resolve documentation gaps and timing issues.
- Contribute to procedure updates and control enhancements focused on accuracy, turnaround time, and risk containment.
Qualifications
- 3+ years in mortgage warehouse lending and/or related industry experience (e.g., construction draw administration or mortgage operations with draw funding responsibilities); title/closing or loan administration experience preferred.
- Skilled in reviewing 1003s, appraisals, closing disclosures, draw requests, inspections, and CO/Completion documentation.
- Ability to validate inspection percent complete, reconcile to funded-to-date, and apply FHB advance limits.
- Strong documentation review and reconciliation skills; high accuracy in calculations and recordkeeping.
- Proficiency with warehouse/LOS platforms and Excel for tracking, reconciliations, and reporting.
- Clear written and verbal communication with internal associates and clients; able to manage volume and deadlines.
- Bachelor’s degree preferred.
