Across a few different teams at Zillow, my work has always been Valuation-focused, building and maintaining automated valuation models (AVMs) and sub-component models of AVMs while optimizing for accuracy but keeping a focused eye on maintaining explainability.Project experience ranges from tackling multiple 'green field' problems by researching, building and deploying new models, while also working to maintain, monitor, and improve existing models to protect against model degradation, bugs, upstream data issues, and real world real estate market extreme events.Specific business problems I have worked to solve by leveraging data science and machine learning techniques include but are not limited to:- Recommending comparable homes.- Converting market-specific valuation models to globally trained models, allowing us to combat cold-start in new markets and leverage valuation inter-market dynamics. - Developing a model that produces a synthetic Real Estate comparative market analysis (CMA) to provide an fully explainable.- Re-designing and implementing a new training of an ensemble model from out-of-fold (OOF) to out-of-time (OOT) in order to combat model bias effects from unprecedented home price appreciation.- Adding listing description features to the Zestimate model.- Debugging outlier Zestimate cases and recommending solutions.Languages: Python, SQL, R, PySparkDeveloper Tools: Git/Gitlab, AWS, Airflow, Docker, JupyterLab/Notebooks, Sagemaker, KubeflowModel Experience: tree models, clustering, regularized regression, multivariate analysis, Hypothesis testing, NLP, ensembling, neural nets (PyTorch)

Proxy records are time-series that reflect changes in climate, typically over thousands or millions of years at course and irregular time steps. Although proxy records are ubiquitous globally, they are irregularly sampled in both time and space, and as such make it difficult to use in combination to robustly infer large-scale patterns of climate change. My postdoc focused mainly on developing multi-proxy statistical methods focused on marine-based proxy records from the Pacific and the Atlantic to estimate full gridded fields of climate variables over time and how those fields changed through the Holocene (past 10,000 years).

• Developed and delivered hands-on STEM curricula to 3rd-5th grade students.• Performed STEM outreach for women, minorities, and under-represented students.• Participated in National Science Foundation research to evaluate the impact of incorporating STEM curricula in K-12 classrooms.• Underwent frequent pedagogical training.

• Developed and taught original curricula to a class of ~140 students (Spring 2012) and ~ 100 students (Spring 2013)• Course abstract: Water is ubiquitous and forms the foundation for life on earth. It transports energy throughout the atmosphere regulating climate and climate change, carves and erodes the surface of the earth, and serves as one of our most precious and limited natural resources. The objective of this course is to provide an understanding of the complexity and importance of the movement, distribution, and quality of water, while emphasizing an application to engineering practices.• Topics included: general hydrological cycle, applied statistics and probability, precipitation and interceptive processes, evapotranspiration, infiltration, streamflow, runoff, extreme events, hydrographs, and select hydrology software.