The Hartford seeks to build high-performing, empathetic analytics teams and deliver smarter product pricing decisions, specifically focusing on the Auto Line within the Personal Insurance Loss and Demand Modeling Team.
Requirements
- Comfort with generalized linear models and related techniques (e.g. ridge and lasso regression, GAMs) and the math behind them.
- Hands-on experience with Python and a track record of delivering clean, literate code in a reproducible environment.
- Experience using GitHub for version control, documentation, code collaboration, and technical project management
- Experience building modeling solutions in cloud-native environments, such as SageMaker, a plus
- Experience with machine learning techniques (e.g. gradient boosted trees) and MLOps frameworks
- Experience with Agile framework
- Experience with data collection, analysis, and modeling
Responsibilities
- End-to-end project leadership of predictive modeling solutions that support the auto class plan.
- Test new data sources and modeling approaches to boost product performance. Engineer features to uncover additional signal for our models.
- Support model implementations with state Departments of Insurance (DOI), serving as a subject matter expert on the models you build and responding to DOI queries.
- Maintain a portfolio of repositories and scripts and suggest improvements to workflows. Assist in the modernization of model deployments using our cloud-based MLOps tools.
- Develop and maintain analytic packages to facilitate data collection, analysis, and modeling.
- Ensure analytic reproducibility, conduct code reviews, and maintain git hygiene. Promote good practices in literate, navigable, and unit-tested code.
- Stay up to date on machine learning techniques, apply them to your work, and share them with your team and broader data science community.
Other
- Master's degree in Statistics, Data Science, Mathematics, Actuarial Science, Economics, Engineering, Physics, or similar analytical field.
- At least 5 years in the insurance industry, preferably in loss cost modeling or personal insurance products.
- Able to independently research solutions, identify when you are stuck, and seek help when needed.
- Demonstrated experience with listening to collaborators and brainstorming outcomes that address their business problems.
- Candidate must be authorized to work in the US without company sponsorship.
