Senior Applied Scientist

The Team

Upstart’s Servicing team supports borrowers after loan origination and works to improve outcomes for investors and borrowers. The team is building data-informed approaches across areas such as collections, loan assistance, payment experiences, and borrower communications, with meaningful opportunity to apply machine learning to problems that have often been addressed manually.

As a Senior Applied Scientist on the Servicing team, you will build predictive and causal models and translate their outputs into decisions that improve how loans are serviced. You’ll partner with Product, Engineering, and Servicing and Collections Operations to evaluate interventions and shape data-informed approaches to payment plans, loan modifications, communications, and autopay.

How you’ll make an impact

• Develop and evaluate machine-learning models that predict borrower repayment behavior and other servicing outcomes.
• Apply causal inference and experimentation to measure the impact of servicing interventions, including loan modifications and payment assistance.
• Translate model predictions and treatment-effect estimates into practical decision algorithms and actions.
• Partner with Product and Engineering on payment experiences, autopay, and borrower communications, and with Servicing and Collections Operations on collections strategy and agent-performance insights.
• Identify and develop machine-learning opportunities in servicing areas that have historically relied on manual optimization.
• Explain analytical choices, tradeoffs, and findings to technical and non-technical partners to inform high-stakes operational and product decisions.

Minimum Qualifications

• Graduate degree in mathematics, applied mathematics, statistics, physics, econometrics, operations research, computer science, or a related quantitative field.
• Experience applying causal inference methods and experimentation to real-world decisions in an applied science or data science context.
• Applied statistics, probability, and experimental-design expertise used to build or evaluate models or interventions.
• Production-quality coding experience in Python, including data manipulation and algorithm implementation.
• 3+ years of experience conducting machine-learning, statistical-modeling, or applied quantitative research in an academic or industry setting.

Preferred Qualifications

• PhD in a quantitative field.
• Experience with payment products or payment experiences, such as autopay, recurring payments, or payment plans.
• Experience with causal machine-learning methods or treatment-effect modeling.
• Experience with supervised learning, model evaluation, and feature engineering, especially with large or complex datasets.
• Experience taking work from exploratory analysis and experimentation through production model development while maintaining analytical rigor and model reliability.
• Ability to translate broad servicing or business problems into structured research questions and communicate complex findings as actionable recommendations.

Position location

This role is available in the following locations: Remote — United States.

Travel requirements

As a digital first company, the majority of your work can be accomplished remotely. The majority of our employees can live and work anywhere in the U.S or Canada (outside of Quebec) but are expected to spend high quality time in-person collaborating via regular onsites and in-person meetings. The onsite cadence varies depending on the team and role; most teams meet once or twice per quarter for 2-4 consecutive days at a time.