Upstart is looking to ensure their AI lending models are fair and explainable to maintain compliance, robustness, and trust with stakeholders.
Requirements
- Programming skills in Python
- Full-stack expertise with all steps of the modeling process from ideation to productionalizing code; OR deep expertise in either statistical modeling or machine learning
- Experience and interest in ML explainability or fairness; OR demonstration of high potential to learn quickly and contribute to these areas.
- Proficiency in a broad array of mathematical, statistical learning and machine learning concepts and applications
- Detailed understanding of building good technical solutions; ability to convert ideas into testable hypotheses and/or next steps
- Ability to meet team standards with respect to quality and velocity with minimal help or direction from peers
- For more complex tasks, ability to break down problems into smaller tasks, and use them to estimate the timeline of executing a larger initiative.
Responsibilities
- researching and implementing industry leading methodologies for explaining complex machine learning models
- evaluating their fairness in the lending space
- mitigating any potential fairness disparities via less discriminatory alternative searches
- strive to understand, at a deep level, the dynamics of our models in order to properly inform business decisions
- building, enabling and evaluating our core production models
- developing new products
- identifying novel approaches to core models
Other
- Strong academic credentials with a master's degree in statistics, mathematics, economics, physics, operations research, computer science, or other quantitative areas of study; Ph.D. preferred
- Ability to meet team standards with respect to quality and velocity with minimal help or direction from peers
- Proactively communicate to stakeholders and teammates with ability to contextualize technical results
- Strong sense of intellectual curiosity balanced with humility, drive and teamwork
- Numerically-savvy and smart with ability to operate at a speedy pace
- Enthusiasm for and alignment with Upstart’s mission and values
- 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.
