SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 177,000 - 230,000 / annual
Adyen is seeking a Machine Learning Scientist to join the Personalize Intelligence team in Chicago, focused on checkout personalization for payments. The role centers on solving a core multi-objective optimization problem: determining which payment methods to display to shoppers and in what order, while balancing conversion maximization against transaction costs and fraud exposure.
You will work with hundreds of millions of annual payment transactions, providing exceptional signal volume, rapid feedback loops, and direct paths from algorithmic iteration to merchant value. The team operates in a causal inference-heavy environment where counterfactual reasoning, learning from logged feedback (contextual bandits), and robust offline-to-online evaluation are daily necessities.
Key Responsibilities:
- Analyze large-scale payment and behavioral datasets to uncover patterns, failure modes, and improvement opportunities
- Formulate, develop, and benchmark machine learning algorithms for multi-objective ranking, contextual bandits, and decisioning under uncertainty
- Design robust offline validation frameworks, counterfactual evaluation pipelines, and online A/B tests to isolate true treatment effects from observational biases
- Deconstruct experiment outcomes beyond top-line metrics, providing quantitative explanations of model behavior and trade-off impacts on merchant margin and conversion
- Write clean, modular Python code to implement models end-to-end, benchmark against baselines, and guide production integration with engineers
- Partner cross-functionally with product managers and domain specialists to translate business objectives into well-posed ML formulations and communicate findings to diverse audiences
Requirements:
- 4+ years of professional experience as a Machine Learning Scientist, Data Scientist, or Quantitative Researcher with proven track record developing models for real-world decisioning problems
- Strong theoretical grounding in applied statistics, probability, predictive modeling, and experiment design; natural thinking about selection bias, counterfactuals, and multi-objective optimization
- Excellence in exploratory data analysis, error analysis, and metric design; ability to diagnose model failures and identify incremental gains
- Solid Python coding skills (SQL a plus); comfortable implementing models end-to-end using core libraries (pandas, NumPy, scikit-learn, LightGBM/XGBoost, PyTorch)
- Clear communication skills to articulate model intuition, assumptions, trade-offs, and translate empirical metrics into actionable business context
Nice to Have:
- Experience with ranking, recommendation systems, contextual bandits, or reinforcement learning
- Exposure to causal inference and observational data techniques (uplift modeling, inverse propensity weighting, survival analysis)
- Practical familiarity with containerization (Docker) or experiment tracking tools (MLflow, Weights & Biases)
- Experience handling tabular or time-series data at scale (Polars, PySpark, Trino)
