Our data science team makes sense of diverse datasets related to key business problems. We’re responsible for the full stack of data analytics, from querying or procuring data, cleaning data, feature generation, problem formulation, error/success metric choice, machine learning/predictive model building, and translation of results into business action. The final product can range from reporting that will be used to advise strategic planning, to recommendation of a specific action, to handing off a predictive model for deployment in production. All of this is done to make our various teams and processes more efficient and to price our products more accurately.
Root is seeking a data scientist to help our team add business value by leveraging vast quantities of data and building more accurate predictive models. Examples of some of the rich data sources we work with include smartphone telematics sensor data, marketing campaign results, and unstructured data from claims notes, phone calls, etc. The ideal candidate will have experience putting well-defined parameters around potentially ambiguous problems, evaluating success based on business-important criteria, and applying modern machine learning techniques to large datasets. This may include querying/scraping/otherwise procuring datasets, cleaning/munging, signal processing, feature engineering, model choice/evaluation, and validation using reproducible methods and scalable architecture. Particular attention will also be paid to the interface between R&D environments and productionizing models. Furthermore, the candidate will need the ability to learn new tools on the fly and be adaptable to changing requirements.
- Mine text data for use in classification/predicting business outcomes.
- Build anomaly detection models to determine fraudulent activity, unexpected trends in data, etc.
- Apply machine learning to image classification for key business applications.
- Identify, evaluate, and productionize new data sources (e.g. geospatial data, web scraping, cell phone sensors).
- Predictive modeling related to improving insurance pricing.
- Strong programming skills, especially R and/or Python
- Demonstrated experience in building, validating, and leveraging machine learning models
- Demonstrated skill with data mining, data munging, coping with missing / corrupt / unstructured data
- 1+ years of industry experience building predictive models OR graduate-level research in a relevant area
- Master's degree or Ph.D. in mathematics, physical sciences, engineering, or other technical discipline
- 3+ years of industry experience building predictive models
- Experience using big data tools (e.g. Hadoop, Spark) and cloud computing (AWS preferred)
- Experience with version control (git/GitHub preferred)
- Advanced knowledge of physics, linear algebra, probability/statistics
- Experience building insurance pricing models
Why join Root? Because we believe the best way to move the world forward is the relentless pursuit of ideas. We build things that ought to exist, but would not exist, if we were not building them. We do this through immense dedication to our craft, an unwavering focus on our fellow human beings, and a maniacal obsession with the quality of our creations. Problems stick in our craw, and we don’t let them go until we solve them. We believe in the scientific process and know that experimentation is the best way to find truth. We don’t hold ideas down no matter how crazy they may seem. We foster them, nurture them, and unceasingly test them. If you are the kind of person that loves tackling big problems with a collaborative group of people until you find solutions, you belong here.
We believe our customers deserve insurance that is:
Fair. We lean on data, not demographic labels.
Affordable. We base rates on good driver performance.
Personal. We give our customers the power to affect their rates.
Easy. We create an intuitive experience.
Accessible. We make information clear.
Beyond. We are never, ever satisfied with the status quo.