About LoanLight
Non-Agency lending is a vital part of the modern mortgage landscape, eclipsing $500B in annual originations and serving borrowers who fall outside the tight constraints of Fannie Mae and Freddie Mac agency guidelines. These borrowers may be self-employed, real estate investors, foreign nationals, or individuals with non-traditional income streams—people who are financially capable but don’t fit inside an agency box. However, unlike agency lending, these loans lack a standard automated underwriting system—resulting in no shared underwriting language and making it difficult to create a scalable, verifiable asset class.
Enter LoanLight: the first AI-native platform purpose-built for non-agency and non-QM lending — transforming how loan files are prepared, validated, and underwritten. By combining deep mortgage expertise with purpose-built AI, we help lenders move faster, investors buy with confidence, and borrowers get to “yes” faster.
Why LoanLight
- Unfair distribution advantage. Stackpoint—our first investor and the leading builder of vertical AI real estate software—gives us instant access to customers, design partners, and capital.
- Experienced founding team. CEO is a repeat fintech founder and product leader that sold his previous company to JG Wentworth in 2024, and led non‑QM diligence at Setpoint (Series B, a16z‑backed).
- Strong Capital Support. We have already raised $1.5M from early customers and advisors, and we have strong investor demand when we choose to raise additional capital. Initial product in market, real traction.
- Massive market, growing fast: Non-QM is the fastest growing segment in residential real estate, and the cost to originate a loan has increased over 30% in the last 2 years. Our agentic AI solution is replacing the $10B spent every year in labor costs to manually underwrite these loans.
- True influence from day one. You’ll shape strategy, product, culture, and execution from the beginning.
About the Role
We’re looking for a Senior Software Engineer, Applied AI to build production-grade AI systems that solve real, high-stakes problems. This role is for an engineer who loves turning messy, ambiguous inputs into reliable, scalable software—using modern ML, LLMs, and traditional systems engineering together.
You’ll work at the intersection of AI, data, and core product, owning systems end-to-end: from model integration and orchestration to APIs, workflows, and reliability in production.
This is not a research role. This is a builder role.
Key Responsibilities
- Full-Stack Feature Delivery
- Architect, develop, test, deploy, and monitor new product capabilities in short, iterative cycles.
- Balance rapid experimentation to inform product discovery with long-term code health and maintainability.
- Agentic AI Workflows
- Translate complex lending workflows into multi-agent systems with LangGraph (and related tooling), handling orchestration, classification, data extraction, validation, and decisioning logic.
- Develop evaluation frameworks and monitoring systems to assess agent performance – both pre-release and once in production.
- Make pragmatic tradeoffs between automation, human-in-the-loop, and reliability