Senior Data Scientist - Insurance Modeling

Pie's mission is to empower small businesses to thrive by making commercial insurance affordable and as easy as pie. We leverage technology to transform how small businesses buy and experience commercial insurance. Like our small business customers, we are a diverse team of builders, dreamers, and entrepreneurs who are driven by core values and operating principles that guide every decision we make.

You will work to establish Pie as the preeminent commercial insurance among small business owners by establishing a best in class data analytics as a AI Data Scientists in this startup environment. You will work with Pie’s Data Science, data and ML engineering, and Product teams to conceptualize, build, and enhance data-driven AI/ML solutions to address various risk and underwriting challenges. You will have the opportunity to impact mission-critical functions by leveraging AI/ML, and see the fruits of your work in action. You will explore the frontiers of explainable machine learning, leveraging advanced supervised/ unsupervised/semi-supervised ML algorithms to build more elegant pricing and risk solutions, construct novel features, and build automated capabilities. Ultimately, you will make AI/ML a key competency, and as easy as Pie.

How You’ll Do It

Working collaboratively with our Product, data engineering, and MLOps teams, you will be actively involved in the entire Model development lifecycle from conceptualization to deployment. You will help conceptualize, design, generate and test hypotheses, construct features, build and validate various pricing, underwriting, and claims models. You will leverage your deep-learning and NLP skills to develop better predictors based on tabular and text data from internal and external sources. 

• Enhance and reinvent the next generation of risk (frequency, severity, LR) and pricing models, with strong emphasis on model robustness 
• Working with business partners, design and build AI-ML solutions in claims, underwriting (UW), customer behaviors use-cases
• Build demand elasticity models, assess the impact on key business metrics of rate changes for different subpopulations, and make recommendations on path forward
• Conduct post-hoc model diagnostics and build interpretability reasons using ML methods
• Monitor and evaluate the performance of various models; detect and come up with mitigation strategies for addressing performance degradation
• Leverage experimentation techniques to construct the best overlays for relevant risk models 
• Monitor relevant KPIs, and develop automation process for revising risk overlays 
• Build new high-signal insightful features, analyzing a diverse set of internal and external data, and leveraging leverage deep …