Assistant Director, Data Science (STP)

Liberty Mutual Insurance•Boston, MA

3d•Hybrid

About The Position

This role is tasked with reengineering the entire modeling pipeline within the Data Science Tools team of Infrastructure of USRM Data Science. The primary goal is to architect and construct modeling tools and pipelines to significantly enhance the ease of iteration, evaluation, and deployment of models, thereby boosting the productivity of the entire USDS department. The position involves building tools and processes to empower other data scientists. A strong understanding of insurance data science, particularly insurance pricing and modeling processes, is essential. The ideal candidate is proactive, highly technical, and driven by a desire to help others and innovate. The role requires domestic (U.S.) travel up to 10% and permits telecommuting up to 60%.

Requirements

Bachelor’s degree (or foreign equivalent) in Computer Science or Data Science or related field and 5 years of experience in the job offered or in a related occupation, OR Master’s degree (or foreign equivalent) in Computer Science or Data Science or related field and 4 years of experience in the job offered or in a related occupation.
Demonstrable experience with insurance pricing and underwriting.
Demonstrable experience with insurance statistical modeling in Emblem, including input data creation and output model formats.
Demonstrable experience with writing Python and SQL with production level ML and Dev Ops code standards, including modular and tested code.
Demonstrable experience with version control using git.
Demonstrable experience with performing open-source Python package development with GitHub, including GitHub projects and Issues, GitHub Actions, Pull Requests, automated linting and formatting, and test-driven development.
Demonstrated understanding of the data science modeling process from dataset creation through model deployment, including feature engineering and statistical modeling.
Demonstrable experience with statistical modeling including GLMs, GBMs, Random Forest, clustering, and neural networks.
Demonstrable experience with modeling pipelines using Airflow, Luigi or Celery with model versioning in MLFlow.
Demonstrable experience with AWS cloud infrastructure (EC2, EMR and S3).
Demonstrable experience with the current state of the Python open-source ecosystem packages including polars, pandas, h2o, sklearn, lightgbm, and pydantic.
Ability to evaluate the newest Python packages for inclusion into our tools.
Demonstrated ability to operate within a large organization and move the company forward with new technological solutions.
Ability to communicate tooling best practices and tool updates via presentations, email and posts.
Ability to train and mentor data scientists on tooling best practices and tool updates.
Proven ability to debug multi-layered problems through a suite of tools.
Demonstrated ability to gain trust of data scientists and partner with them to solve their coding problems.

Responsibilities

Reengineer the approach to the entire modeling pipeline within the Data Science Tools team of Infrastructure of USRM Data Science.
Architect and construct modeling tools and pipelines to enhance the ease of iteration, evaluation, and deployment of models.
Massively enhance the productivity of the entire USDS department.
Build tools and processes to empower other people.