SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 170,000 - 220,000 / annual
Arlo is rebuilding health insurance for small businesses by redirecting premium dollars toward actual care rather than administrative overhead, fraud, and middlemen. The company uses AI across underwriting, operations, clinical programs, and member experience to build a more efficient insurer. Operating at meaningful scale with hundreds of millions in premiums, tens of thousands of members, and backing from Upfront Ventures, 8VC, and General Catalyst.
As Senior Data Scientist for Underwriting, you will own Arlo's core underwriting model—the system that estimates individual member risk and enables competitive group pricing. You'll work with billions of claims across tens of millions of patients to identify signals in claims history that predict future medical costs, develop robust pricing strategies, and continuously deploy measurable improvements based on real-world outcomes.
Key responsibilities include: evaluating the existing underwriting model to identify gaps in risk estimates across cohorts, conditions, and claims patterns; developing a prioritized roadmap of model and feature improvements that balance competitiveness with profitability; building features that capture full member risk while accounting for dataset bias and cost variance; experimenting with ML architectures that balance efficacy, generalizability, and interpretability; and proving the lift of every model change through backtesting before deployment.
You'll sit alongside data scientists and actuaries, thinking through issues beyond point estimates—including data blindness, variability, and risk assessment for quote issuance. You'll own the model end-to-end while collaborating with ML engineers to ensure ideas can be tested and deployed at scale. This is a hands-on role focused on shipping production models that directly impact win rates, book size, and company performance.
Required: 5+ years building predictive models in production; deep Python and SQL proficiency with experience processing large datasets using Spark; track record of owning ambiguous problems end-to-end; direct healthcare data experience; strong feature engineering and model validation instincts; and clear communication skills. Nice-to-have: claims data familiarity, Series C or earlier startup experience, underwriting/actuarial background, or ML engineering experience.
