Assistant Director, Data Science, Solaria Labs
Quick Summary
Description Solaria Labs is Liberty Mutual’s innovation team. We partner with business leaders across the enterprise to incubate new products, services and experiences.
Solaria Labs is Liberty Mutual’s innovation team. We partner with business leaders across the enterprise to incubate new products, services and experiences. We prove ideas out quickly and transition successful solutions into the business.
The Applied Science team leads Solaria’s AI experimentation and delivery. We apply frontier AI to real business problems, including agentic workflows, enterprise knowledge systems, GenAI evaluation, and predictive modeling.
What sets this role apart:
• Breadth: You’ll work across commercial and retail insurance and enterprise functions such as Finance, Talent, and Investments.
• Visibility: You’ll work at the intersection of senior business and technology leaders.
• Research: Our world-class partnership with MIT keeps our work at the forefront AI.
• Lasting impact: You’ll build reusable capabilities, such as evaluation frameworks, tools and knowledge layers, that extend beyond any single project.
Discover more at www.solarialabs.com
Responsibilities
~1 min read• Design and build AI solutions across predictive ML, NLP, knowledge systems/GenBI, and agentic workflows for high-priority incubation programs
• Design evaluation frameworks (golden datasets, validation, monitoring) that make GenAI and agentic systems trustworthy and measurable
• Build durable, reusable components (MCP servers/tools, agent skills, knowledge layers, evaluation suites) that outlast any single program or platform
• Advise on when to build custom solutions versus leverage enterprise AI tools (e.g., Claude, ChatGPT Enterprise)
• Serve as technical lead on complex programs, shaping problem framing, architecture, and success metrics with PM, Engineering, UX, and subject matter experts
• Translate results into clear recommendations and trade-offs for senior and executive audiences
• Contribute to the Applied Science research agenda, including academic partnerships such as MIT
• Mentor and guide Co-Op and IEP talent on research and program work
• Move fluidly across programs as the portfolio evolves
Requirements
~1 min read• Builds durable relationships across functions and business units
• Strong foundation in statistical modeling and machine learning, plus hands-on experience building LLM-based applications
• Experience designing evaluation approaches for ML or GenAI systems
• Proven ability to own ambiguous problems end to end, from framing through delivery
• Ability to communicate complex ideas clearly and influence decisions with business and technical leaders
• Competencies typically acquired through a Ph.D. degree (in Statistics, Mathematics, Economics, Actuarial Science or other scientific field of study) and a minimum of 2 years of relevant experience, a Master's degree (scientific field of study) and a minimum of 4 years of relevant experience, or may be acquired through a Bachelor's degree (scientific field of study) and a minimum of 5+ years of relevant experience.
• Ability to work EST hours
• Experience with agentic frameworks (e.g., LangGraph) and tool-use patterns (MCP, function calling)
• Experience with RAG, semantic layers, or enterprise knowledge systems
• Familiarity with cloud data platforms (e.g., Snowflake) and MLOps/LLMOps practices.
• Experience with graph or anomaly detection methods
• Experience in research settings or academic collaborations
• Curiosity, fast learning, and comfort with shifting priorities
What We Offer
~1 min readLocation & Eligibility
About this job
- Who can apply
- North America
- Posted
- 734 days ago
Listing Details
- Posted
- November 17, 2024
- First seen
- November 17, 2026
- Last seen
- November 22, 2026
Posting Health
- Days active
- 4
- Repost count
- 0
- Trust Level
- 18%
- Scored at
- November 22, 2026
Signal breakdown
Browse Similar Jobs
Stay ahead of the market
Get the latest job openings, salary trends, and hiring insights delivered to your inbox every week.
No spam. Unsubscribe at any time.