Role Overview
Stripe's Growth Platform builds machine learning systems that help businesses discover and use Stripe products across Dashboard, email, onboarding, documentation, and AI agent interfaces. The team combines business context with recommendation, ranking, contextual bandit, and agent-based models, plus the data and evaluation systems behind them.
You will build and operate production ML systems, owning work from problem definition and feature development through training, evaluation, deployment, monitoring, and iteration. You will partner with data scientists, engineers, and product teams to turn model improvements into measurable user and business outcomes.
Responsibilities
- Design, train, evaluate, deploy, and maintain recommendation, ranking, and personalized action models.
- Improve contextual bandit and policy-learning approaches, including exploration, reward design, and context adaptation.
- Build agent-based recommendations that use business context to identify relevant products and integration options.
- Develop reliable data and feature pipelines for training and inference, improving freshness, quality, and consistency.
- Build reusable tooling for evaluation, retraining, and safe rollout.
- Own production ML quality, including tested code, monitoring, failure investigation, reliability, latency, and cost.
- Design and analyze online experiments, connecting offline evaluation to adoption and incremental impact with guardrails.
- Partner with product engineering and ML infrastructure teams to integrate models and improve shared capabilities.
Minimum Requirements
- 3+ years of industry experience in machine learning engineering, software engineering, or applied data science, with hands-on experience building and shipping ML models in production.
- Strong programming skills in Python and experience writing maintainable, tested production code.
- Practical experience designing, training, and evaluating ML models using frameworks such as PyTorch, TensorFlow, XGBoost, or scikit-learn.
- Experience building data or feature pipelines, proficiency in SQL, and familiarity with distributed data processing tools such as Spark or PySpark.
- Strong understanding of statistics, model evaluation, and experimentation, including recognizing data leakage and distinguishing offline model improvements from business impact.