Andrew Schechtman-Rook Email & Phone Number

Vienna, VA, US

United States

Washington D.C. Metro Area

• Deployed the first cloud-based credit card underwriting model in the company via a dockerizedPython API, with an estimated incremental value of $35MM per year.
• Core maintainer for the longest-lived, most successful internal data science tool in the com-pany, used in production models by dozens of data scientists across lines of business.
• Owned the design, implementation, and upkeep of the DS Technical Interview, a key interviewgiven to every data science candidate in the US.
• Guided ongoing development of the core credit card valuations model scoring platform, delivering regular releases of new and updated models while improving the robustness andmaintainability of platform infrastructure.
• Led technical development of model monitoring tools, mentoring three junior data scientiststo deliver a maintainable package on time and to spec.

Washington D.C. Metro Area

• Built a Python library to facilitate the creation of robust, production-ready data analysis pipelines, saving hundreds of hours of R&D effort across multiple production models.
• Led development of a prototype automated machine learning model deployment framework for cloud-based applications, influencing the development direction for the company-wide credit card application processing platform.

Washington D.C. Metro Area

• Implemented a novel approach to deliver internal technical trainings, providing over 5000 hours of classes with no instructors.
• Programmed and deployed an interactive course completion dashboard using Flask and dc.js to provide progress reports to individual students as well as company leadership.
• Engineered a credit card payment simulator, allowing a product team to gain actionableinsights without expensive pilots.

Greater New York City Area

• Trained a multiclass classification model on historical airline on-time arrival data to predict flight delay times.
• Designed a web interface to simplify itinerary input and intuitively display results.

Madison, Wisconsin Area

• Devised metrics to improve correspondence between numerical models and astronomical data. Implemented in highly optimized Python, was able to refine agreement by up to 20% with minimal increase in computation time.
• Built a fast Voronoi Tessellation algorithm to adaptively bin images, preserving spatial resolution while maximizing signal in images with over one million pixels.
• Constructed visualization tools using Matplotlib to aid data processing efforts.
• Trained and mentored undergraduate and graduate students in programming, data analysis, and statistical methods.

• Developed non-linear Levenberg-Marquardt chi-squared fitting algorithms using a combination of Python and C++ to constrain models of spiral galaxies to data.
• Employed on-campus distributed computing resources to perform large-scale modeling in parallel, using over 20 years of computer time in one month.
• Assembled a hybrid C++/Python processing pipeline to clean, register, and mosaic hundreds of high-resolution images with minimal user intervention, resulting in a factor of 10+ increase in analysis precision.
• Created a genetic algorithm in C++ to efficiently fit galaxy models with unusually large numbers of free parameters to high-resolution images.
• Utilized frequency-domain analysis to understand the spatial distribution of galactic structure.
• Computed descriptive statistics about 200+ astronomical objects from raw survey data automatically via custom-built analysis software blending C++ and shell scripting programs.

Madison, Wisconsin Area

• Taught six discussion sections of an introductory undergraduate astronomy course.
• Prepared engaging lesson plans, including interactive demonstrations and group problem-solving activities.

• Performed novel statistical analyses on publicly available NFL data.
• Mined a play-by-play database containing over 500,000 records across dozens of tables for complex relationships between individual players as well as teams.
• Developed custom software and visualization tools using MySQL and Python to efficiently examine thousands of plays.
• Explained findings in a manner accessible to all audiences through both written posts and evocative figures.

• Designed and executed statistical analyses of simulated galaxy clusters to optimize strategy for future data acquisition.
• Used nearest-neighbor and regression analysis to compute a transformation between astronomical filter systems.

• Maintained existing equipment and computers for introductory physics labs.
• Developed and built components for new labs.
• Upgraded hardware and software for over 20 lab computers.

Doctor Of Philosophy (Phd), Astronomy

Master Of Science (Ms), Astronomy

Bachelor Of Science (Bs), Astronomy