Senior Software Engineer, Personalization & ML

Upstart Upstart · Fintech · Remote · Engineering

Senior Software Engineer role focused on personalization and ML within Upstart's Servicing Engineering team. The role involves embedding ML models into product and operational workflows, advancing the experimentation platform, improving strategy performance evaluation, and scaling model-driven decisioning through feature pipelines and data integrations. The goal is to personalize borrower experiences and improve servicing decisions using AI.

What you'd actually do

  1. Improve how Servicing decisions are made by embedding machine learning models into product and operational workflows.
  2. Enable faster learning and safer iteration by advancing our experimentation platform and improving how we evaluate strategy performance.
  3. Increase the effectiveness of personalization strategies by designing and running controlled experiments that translate into measurable improvements.
  4. Scale model-driven decisioning through resilient feature pipelines and real-time data integrations.
  5. Define clear metrics and guardrails to ensure ML-powered systems remain measurable, explainable, and compliant as they shape more Servicing decisions.

Skills

Required

  • Bachelor’s degree in Computer Science, Engineering, or Mathematics, or a related field (or its equivalent) + 4 years of experience
  • Experience owning delivery of ML-powered features from design through production deployment and measurement.
  • Hands on experience designing or contributing to experimentation systems, including running controlled experiments in live environments.
  • Experience building and maintaining data processing systems or pipelines that support model-driven decisioning.

Nice to have

  • Experience with building or scaling ML-powered ranking, personalization, or recommendation systems in production environments.
  • Applied advanced experimentation methods beyond standard A/B testing.
  • Demonstrated incorporation of fairness, explainability, or governance considerations into ML-powered decision systems.
  • Led technical design decisions for distributed systems supporting ML-driven workflows.

What the JD emphasized

  • ML-powered features from design through production deployment and measurement
  • experimentation systems
  • controlled experiments in live environments
  • data processing systems or pipelines that support model-driven decisioning
  • ML-powered ranking, personalization, or recommendation systems in production environments
  • fairness, explainability, or governance considerations into ML-powered decision systems

Other signals

  • productionizing model outputs
  • scale model-driven decisioning
  • ML-powered features from design through production deployment and measurement
  • experimentation platform
  • personalization strategies