Modeling & Statistics
• Working knowledge of GLMs and GBMs in Python, with an understanding of when each approach is appropriate (e.g., GBMs for exploration and interaction detection; GLMs for interpretability and implementation readiness).
• Ability to assess model stability, identify overfitting, and communicate results and tradeoffs clearly.
• Familiarity with ML lifecycle best practices including documentation, version control (GitHub), and experiment tracking (e.g., MLflow).
Data & Vendor Support
• Experience with data validation, quality checks, and working across multiple data sources simultaneously.
• Comfortable engaging with external vendors or data providers to ask clarifying questions and resolve data issues.
Business Communication
• Ability to present analytical findings clearly to both technical and non-technical audiences.
• Developing skill in translating model results into actionable recommendations, including communicating uncertainty or limitations honestly.
Technical Foundations
• Bachelor’s or Master’s degree in Computer Science, Mathematics, Data Science, or a closely related discipline.
• Experience in statistical modeling and machine learning using Python (pandas, NumPy, scikit-learn) with strong SQL skills.
• Across the modeling lifecycle: problem framing, experiment design, evaluation, and validation.
• Experience using Git and Unix-based development environments with reproducible analytical workflows.
• Familiarity with model monitoring concepts including drift detection and performance tracking.
• Some exposure to cloud-based platforms (Vertex AI, SageMaker, or Azure ML) is a plus.
• Familiarity with enterprise governance expectations including compliance, privacy, and model documentation standards.
________________________________________
Nice to Have
• Experience in regulated modeling environments, including documentation and approval workflows.
• Familiarity with insurance pricing, segmentation, or rating variables.
• Familiarity with bias/fairness testing and model risk documentation.
• Exposure to generative AI or LLM concepts (RAG, prompt engineering, agentic workflows).