Director, Data Science - Rollout acceleration AI

Liberty Mutual•Boston, MA

About The Position

US Data Science (USDS) is building a new team to use advanced AI to help insurers bring pricing programs to market faster. This highly visible team will build evaluations for AI and agentic systems, measuring their value, accuracy, and risks so high-value applications can be brought into production with confidence. Its work will shape the use of AI in core pricing work for US Retail Markets (USRM). The Director, Data Science will serve as the data science partner for a line of business—Property & Specialty, Auto, or Small Commercial—while collaborating with two other senior data scientists and sharing methods and tools. The role reports into the Rollout AI team in USDS and has a dotted-line relationship with its assigned line of business. The position may be filled at the Assistant Director or Director I level, depending on skills and experience, at manager discretion. Candidates within 50 miles of Boston, MA; Portsmouth, NH; Seattle, WA; Columbus, OH; or Plano, TX will work a hybrid schedule, with two days per week in the office. Otherwise, the role is remote with occasional travel. In-office requirements may depend on candidate location.

Requirements

Hands-on experience building and deploying insurance pricing models.
Strong expertise in model evaluation, statistics, and experimental design.
Strong coding skills, for example in Python, and a track record of building and shipping data science solutions.
Strong, recent hands-on experience with modern LLMs, including agentic coding, building LLMs into applications, and/or RAG.
Strong communication and stakeholder skills, with comfort working across business, delivery, and technology teams.
Resilient and curious; acts in service of the best answer, holds themselves to a high standard, and wants to work with others who do the same.
Broad knowledge of predictive analytic techniques and statistical diagnostics of models.
Advanced knowledge of predictive toolsets and expertise in tool development.
Ability to exchange ideas and convey complex information clearly and concisely.
Ability to establish and build relationships within and outside the organization.
Ability to give effective training and presentations to management and other groups.
Ability to use analysis results to persuade teams or department and senior management to a course of action.
Broad knowledge of business drivers and market context, with a value-driven perspective on work context and impact.
Competencies typically acquired through a Ph.D. in Statistics, Mathematics, Economics, Actuarial Science, or another scientific field and at least 3 years of relevant experience; a master's degree in a scientific field and at least 6 years of relevant experience; or a bachelor's degree in a scientific field and at least 8 years of relevant experience.

Nice To Haves

Experience evaluating generative AI or agentic systems.

Responsibilities

Partner with pricing, delivery, and technology teams in the assigned line of business to identify where AI can accelerate pricing program rollout.
Build evaluations for AI and agentic applications.
Define accuracy metrics and ground truth; generate statistical tests and edge cases; and design experiments, including tests in production.
Measure the value and risk of AI applications, including effort reduced, errors avoided, and time to market gained.
Bring pricing expertise and product sense to AI projects by translating business needs into requirements and evaluation criteria.
Prototype and build evaluation tooling and, over time, AI applications that accelerate rollout.