Data Scientist - Modeling and Analytics

AAA Life Insurance Company•Livonia, MI

22h•Hybrid

About The Position

AAA Life is a respected and trusted American brand that has been focusing on Life Insurance and Annuity Products since 1969. At AAA Life we have over 1.8 million policies where we take pride in earning the trust of our policyholders who understand our promise to be there for them – and their families – when we’re needed most. By joining the AAA Life team, you are joining a company that genuinely cares about helping each other, with a devotion to protect the lives of those around us. We embrace a diverse, equitable, inclusive culture where all associates can feel a sense of belonging and use their unique talents and perspective to influence, innovate, motivate, and thrive.

Requirements

Skilled in cross-functional collaboration, agile methodologies, project management and stakeholder communication.
Advanced training or academic focus in non-parametric statistics, resampling methods, or Bayesian approaches for small sample inference.
Experience applying sequential testing or multi-armed bandit approaches to maximize insights from limited samples in marketing contexts.
Able to effectively communicate and translate complex, technical finding in a candid, clear, concise, and non-technical fashion to all audiences.
Maintain perspective between the big picture and the tactical details.
Remains aligned with the organization’s strategic plan.
Stellar attention to detail, including maintaining accuracy and consistency across a suite of data science assets, keeping documentation up to date, and proactively identifying and addressing any quality concerns.
Self-starter with the ability to identify priorities and focus on items with high business impact.
Ability to present complex analytical findings with persuasiveness and succinctness.
Proficiency in SQL for data manipulation.
Sagemaker or other AI/ML tools for model building.
R or Python for data analysis.
A visualization tool such as PowerBI for quickly assessing model performance.

Nice To Haves

Master’s degree in Statistics, Economics, Mathematics, Data Science, or related field.
Experienced in marketing analytics or customer behavior modeling.
5 to 7 years of experience in data science, including hands-on experience with Machine Learning (e.g., scikit-learn, TensorFlow, PyTorch, DataRobot, Databricks) and Generative Artificial Intelligence.
Experience with automated model deployment and monitoring tools.
Possess outstanding analytical, modeling, problem-solving, and critical-thinking skills.
Experienced with cloud platforms such as AWS, Azure, and Google Cloud.
Familiar with big data technologies (Spark, Hadoop).
Strong knowledge of machine learning algorithms and their applications in automated systems.
Experience with advanced modeling techniques like ensemble methods, time series analysis, and probabilistic modeling.
High proficiency in Python or R for statistical analysis, model development, and process automation.
Proficient+ with SQL for data extraction and manipulation.
Proficiency with data visualization tools (Power BI, Tableau, or similar) and their automation capabilities.

Responsibilities

Build, maintain, and automate models to predict purchase propensity, policy premium, policy lapse/retention, cross-selling, upselling, next best action, and other consumer behaviors using both internal data, census data, appended aggregated data, and macroeconomic data.
Recommend marketing distribution strategies leveraging data and models.
Conduct advanced exploratory data analysis.
Perform model interpretability and explainability analysis.
Leverage specific metrics for model performance evaluation (e.g., precision, recall, F1 score).
Implement A/B testing and experimental design and quantitative benchmarks for model improvement.
Apply data privacy and compliance rules under regulations like GDPR, CCPA.
Apply ethical AI principles.
Apply model fairness and bias mitigation techniques.
Conduct analyses to assess model performance and campaign performance, both against test datasets and actual results once deployed.
Forecast campaign results based on models built and validate forecast against actuals.
Work with marketing data architects and engineers to ensure data is clean, complete, correct, and suitable for modeling using AI/ML platforms.
Develop and maintain data pipelines.
Implement feature engineering techniques.
Find, recommend, and purchase additional data to use in model building.
Proactively identify opportunities for model improvement and need for additional modeling projects.
Maintain clear and organized documentation of data, methodologies, and results.
Implement automation in existing processes to improve overall efficiency.
Perform ad hoc analysis to support Marketing Distribution efforts.
Actively seek out innovation and optimization use cases and experiments that will result in organizational transformation and sales and profit improvements.