• About The Role We are looking for a Senior Data Science Researcher to join Honeycomb's AI team, the team behind underwriting decisions, quote-time risk scoring, and portfolio analytics. In this role, you will own entire research directions: framing the problem, deciding what's worth measuring, designing the approach from first principles, and turning the result into production signal. The work is production-oriented applied research on hard, real-world datasets where standard recipes don't apply and the right method often has to be invented for the problem. What You'll Do - Own open-ended research directions end-to-end - from a vague business question to a deployed, validated signal. - Work across the team's core research areas — risk modeling, environmental and catastrophe modeling, and behavioral modeling - translating complex real-world processes into reliable predictive signal. - Deeply investigate existing production models - understand what they actually learn, where they break, and where the headroom is; identify and ship optimizations grounded in that understanding. - Design approaches from first principles when off-the-shelf methods don't fit the data. - Build rigorous evaluations and own the result in production, not just the notebook. - Document research clearly so findings are reproducible, auditable, and compound over time. Requirements - MSc. or PhD in Statistics, Applied Math, Physics, or a closely related quantitative field. - Real research experience - a track record of driving original investigations end-to-end (academic research, a research-heavy PhD, industry R&D, or equivalent), not only applied ML delivery. - Strong mathematical or statistical problem-solving instincts - able to model a messy real-world system from scratch, not just apply a library. - Production deployment experience: monitoring, CI/CD, data validation, reproducibility. - Ability to independently initiate, plan, and drive entire research directions. - Team player, positive, driven, independent, fast learner.

Advantages

• Depth in classical statistics, causal inference, or applied probability. - Deep learning experience. - Clean coding and repository-maintenance habits.