Upstart is looking to improve its AI lending marketplace by enhancing the performance, efficiency, and reliability of its machine learning models, which are central to its business operations. The company aims to increase approval rates and reduce loss rates while providing a superior digital lending experience.
Requirements
- Programming skills in Python
- Proficiency across machine learning, numerical computing, and software engineering fundamentals
- Foundations in probability and statistics
- Experience building models and conducting statistical or quantitative research
- Experience building ML tooling or solving real‑world ML engineering problems in industry (e.g., data pipelines, evaluation frameworks, training/serving), such as through prior internships
Responsibilities
- Implement algorithmic improvements that boost predictive power, training efficiency, or serving latency
- Design reliable data/feature pipelines
- Automate repeatable workflows that reduce time from research to deployment and monitoring
- Deliver projects that improve model performance, efficiency, latency or reliability
- Write production-grade code, i.e. tested, reviewed and scalable Python
- Communicate findings clearly to get buy-in for recommended next steps
- Work with your mentor to align stakeholders, and drive next steps
Other
- Strong academic credentials with an ongoing bachelor’s or master’s degree in computer science, physics, machine learning, or other quantitative areas of study.
- We require that you are on track to graduate by the summer of 2027 (our internship program is not open to first- and second-year undergraduates).
- Strong sense of intellectual curiosity, humility, drive and teamwork, as well as communication skills.
- PhD studies or post-doctoral research in computer science, physics, machine learning, or other another quantitative field. If you are a PhD student, we require that you are on track to graduate by the summer of 2027.
- The team operates on the East/West coast time zones.
- The majority of our employees can live and work anywhere in the U.S but are encouraged to to still spend high quality time in-person collaborating via regular onsites.
- The in-person sessions’ cadence varies depending on the team and role; most teams meet once or twice per quarter for 2-4 consecutive days at a time.
