Gemini is looking to protect its customers and platform from fraud risk by designing and deploying machine learning models to detect, prevent, and mitigate fraud across its ecosystem.
Requirements
- Strong proficiency in Python and relevant modeling libraries (e.g., scikit-learn, xgboost, TensorFlow, PyTorch) and SQL.
- Experience with data processing and model lifecycle tools such as Databricks, SageMaker, Snowflake, MLflow, or similar.
- Familiarity with orchestration and data pipeline frameworks (e.g., Airflow, Spark).
- 1+ years of experience developing, deploying, and maintaining production-grade ML models, ideally for real-time or large-scale applications.
- 5+ years of experience applying data science and machine learning to financial, payments, or fraud-related problems.
Responsibilities
- Analyze large, complex datasets to identify key fraud indicators and engineer predictive features using internal and external data sources.
- Design, train, and deploy machine learning models to identify and prevent fraud, including payment fraud, account takeovers, and identity abuse.
- Build and maintain end-to-end data and model pipelines for risk scoring, anomaly detection, and behavioral profiling.
- Evaluate model performance through experiments, backtesting, and continuous monitoring to improve capture rates and reduce false positives.
- Stay current on emerging fraud tactics and machine learning approaches to continually evolve Gemini's defenses.
Other
- Bachelors degree in Computer Science, Data Science, Statistics, or a related field.
- Excellent communication skills and the ability to translate complex technical concepts into actionable insights.
- Ability to work in person twice a week at either San Francisco or New York City, NY office.
- Must be eligible to work in the United States
- Must be willing to work in a hybrid work environment with flexible time off
