Data Scientist

About Finosu Finosu is an AI-native consumer loan servicer. What is that? A servicer handles the communications and payment rails for lenders. The servicing industry services over 10 trillion dollars of debt a year. Loan servicing is the perfect application for AI. Borrowers want help on their time, not during call center hours, and AI makes 24/7 support real for the first time. It also makes compliance better, not worse: every interaction follows the rules, every disclosure is delivered, and everything is logged and auditable. What makes us different is how we measure success. We are not judged on generic CSAT scores. We are judged on dollars generated for our lenders, and it turns out the two go together: borrowers who are treated well perform better. We know this problem because we lived it. Our founders built a $200M consumer lender together, owning the end-to-end stack from origination through servicing and collections. We felt every broken part of it firsthand: the vendors that didn't care about outcomes, the compliance gaps, the borrowers who fell through the cracks. So we are building the thing we wished we had. Things to Know We are a start-up and our processes are in flux. Priorities might shift more abruptly than at a more establish company.

• We’re in office, we have found that collaboration especially in a context-rich field is important
• You will own the data architecture for a platform handling millions of dollars of loan balances and millions of LLM-driven borrower interactions.

About the Role We’re looking a data scientist to lead all the data endeavors at Finosu. You’ll use our data to build the best possible servicing experience for borrowers. That means understanding when to reach out, which channel to use, what to say, and where our product could do more to help. Better borrower support leads to stronger payment performance, and we measure success by the dollars we generate for lenders. You’ll also build our ability to predict performance: how much a portfolio is likely to generate, over what period, and with what confidence. Your work will help us set lender expectations, recommend servicing strategies, and guide product decisions. An ideal candidate would come up with hypothesis and create the tooling needed to verify or invalidate their hypothesis. We have a small team of less than 10 people. In this role

• Improve how we communicate. Determine the right timing, frequency, channel, and content for borrower interactions across voice, chat, SMS, email, and direct mail.
• Find gaps in the borrower experience. Use conversations, product usage, and payment data to identify where borrowers get stuck, what support they need, and how our services should improve.
• Own experimentation. Develop hypotheses, build the tooling to test them, and measure whether changes improve borrower outcomes and payment performance. Make practical decisions when data is limited, while being clear about uncertainty.
• Predict performance. In this role, you will build and validate models that forecast payment outcomes across borrowers and portfolios.
• Turn findings into action. Work directly with the founders and engineers to put improvements into production and translate your findings into useful recommendations for lenders.
• Build our data foundations. Define the data, metrics, and systems we need to make better decisions, and identify additional data sources that meaningfully improve our capabilities.
• 30/60/90 • 30 days: understand our data and borrower journeys, establish performance baselines, and identify the highest-impact opportunities to improve servicing.
• 60 days: run experiments on communication and servicing strategies, and build an initial forecasting model to benchmark against actual payment results.
• 90 days: put validated improvements into production and establish a repeatable process for forecasting performance, evaluating accuracy, and making lender recommendations.

Projects people in this role have taken or will take:

• Determine when a call, text, or email assists and when additional outreach is detrimental.
• Identify why borrowers struggle to create or maintain a payment plan, then test changes that help them follow through.
• Test how message content and agent behavior affect borrower understanding and payment outcomes.
• Forecast a portfolio’s payments over time, explain differences between expected and actual results, and recommend changes to its servicing strategy.

About You

• We have a roadmap, but you’ll own your work: propose what to build, challenge priorities when you see a higher-impact path, and help the team ship what best serves our users.
• You have a graduate degree in statistics or a similar field.
• You love measuring things. As we move fast, the solutions to many challenges are hidden in the data.
• You’re product-minded and data-minded. In this role, you’ll make many decisions to run experiments across the company.