Assistant Director, Data Science

Liberty Mutual Insurance•Boston, MA

1d•$161,803 - $197,000•Hybrid

About The Position

Identify and aggregate sources of third-party geospatial data and use geospatial data management techniques to create usable data and useful insights for Property & Casualty insurance modeling. Build out team’s geospatial big data infrastructure using geospatial data science techniques & software such as Snowflake, dbt, and Python. Build geographic risk segmentation models using geographic data science modeling techniques. Apply insurance knowledge to conduct multiple forms of statistical analyses, such as building loss cost models and creating geographic loss data. Translate big geographic data into production-level Property & Casualty insurance geographic pricing factors using sophisticated geographic data science and cartographic techniques for non-technical audiences. Apply spatial data techniques to create usable outputs from geographic risk segmentation models at multiple geographic granularities. Conduct geospatial data science research to enhance geographic pricing sophistication, including leveraging spatial ML techniques. Identify and test hypotheses, ensuring statistical significance, and build predictive models for business application. Enable the business to make clear tradeoffs between and among choices, with a reasonable view of the likely outcomes. Customize analytic solutions to specific client needs. Responsible for larger components of projects of moderate to high complexity. Guide aspects of project design as a technical consultant for the team. Regularly engage with the data science community and participate in cross functional working groups.

Requirements

Geospatial data science methodologies and applications including geospatial big data management and spatial statistics.
Property & Casualty insurance data science techniques including working with exposure, loss, and premium data.
Statistical modeling methodologies and applications including Generalized Linear Models (GLMs) and Gradient Boosting Machines (GBMs).
Insurance predictive modeling experience in third-party software including Emblem and Earnix.
Python data science libraries including pandas, geopandas, and numpy.
SQL based data management, including Snowflake and dbt.
Data analytics and interpretation through exploratory data analysis and spatial data analysis techniques.
Utilizing AWS cloud services including EC2 and S3.
Demonstrated understanding of MLOps standards including unit testing and virtual environments.

Responsibilities

Identify and aggregate sources of third-party geospatial data and use geospatial data management techniques to create usable data and useful insights for Property & Casualty insurance modeling.
Build out team’s geospatial big data infrastructure using geospatial data science techniques & software such as Snowflake, dbt, and Python.
Build geographic risk segmentation models using geographic data science modeling techniques.
Apply insurance knowledge to conduct multiple forms of statistical analyses, such as building loss cost models and creating geographic loss data.
Translate big geographic data into production-level Property & Casualty insurance geographic pricing factors using sophisticated geographic data science and cartographic techniques for non-technical audiences.
Apply spatial data techniques to create usable outputs from geographic risk segmentation models at multiple geographic granularities.
Conduct geospatial data science research to enhance geographic pricing sophistication, including leveraging spatial ML techniques.
Identify and test hypotheses, ensuring statistical significance, and build predictive models for business application.
Enable the business to make clear tradeoffs between and among choices, with a reasonable view of the likely outcomes.
Customize analytic solutions to specific client needs.
Responsible for larger components of projects of moderate to high complexity.
Guide aspects of project design as a technical consultant for the team.
Regularly engage with the data science community and participate in cross functional working groups.