Data Scientist

A leading financial services organization is seeking an experienced Data Scientist to support large-scale life insurance underwriting initiatives. This role is highly focused on mortality risk modeling, traditional machine learning, and statistical modeling—not data engineering or GenAI-heavy work. You’ll work within a production, business-facing environment and collaborate closely with underwriting teams to develop models that guide premium decisions and risk assessments. Key Responsibilities • Build, validate, and maintain machine learning/statistical models to support underwriting decisions. • Analyze structured health, demographic, and insurance data to model mortality risk and premium outcomes. • Translate underwriting and business requirements into technical analytical approaches. • Collaborate with underwriters, analysts, and cross-functional partners to explain model outputs. • Perform EDA, feature engineering, and ongoing model performance evaluations. • Document assumptions, methodologies, and results in a structured manner. • Support model lifecycle activities, including monitoring and refinement. • Work independently while contributing within a broader analytics team. Required Skills Technical Skills • Strong experience with traditional data science and machine learning. • Proficient in Python and SQL. • Hands-on experience with: • Regression modeling • Random Forest • Gradient boosting (XGBoost preferred) • Experience working with Git/GitHub. • Familiarity with standard data science workflows and best practices. • Master’s or PhD in Data Science or related field. Core Competencies • Self-directed, proactive, and able to operate with minimal oversight. • Strong problem-solving and analytical skills. • Able to work within regulated, data-sensitive environments. • Effective communication with technical and non-technical stakeholders. Plusses • Background in life insurance, underwriting, healthcare analytics, or health data. • Exposure to actuarial concepts or risk modeling. • Light experience with AI/GenAI tools (~10–15% usage).