Passionate about optimization, automation, and statistics-driven decision-making.
Highly proficient with a variety of full-stack application development tools that bring analysis to life.
Passionate about optimization, automation, and statistics-driven decision-making.
Highly proficient with a variety of full-stack application development tools that bring analysis to life.
Led the development of a widely-used internal website running on AWS using React and Django. Created and maintained testing suite and documentation. Onboarded cross-functional contributors and socialized adoption.
Created and maintained data "plumbing" automations that connected platforms and services, including: lab information database, project management software, networked lab instruments, and internal website server. Built a QR-based inventory update system to reduce cycle counting need. Fully automated several lab instrument data ingest operations.
Created a primate immunogenicity forecasting and decision-support tool using stochastic modeling to anticipate spontaneous interruptions to animal availability (a widespread challenge in biotech). Fully-automated analysis pipeline makes increasingly precise recommendations available to decision-makers as new assay results are recorded in lab information database.
Deployed software to support: scheduling automation and visualization; highly customized BI analyses; a pipeline for long-running bioinformatics calculations; IP-related data mining; automated "handoffs" across platforms; integrations with lab robots.
Taught "Data and Analytics" to students in the full-time and fully-employed MBA programs as part of PhD teaching requirements.
Built a git deployment pipeline for learning material using JupyterHub. MBA students clicked a "magic link" to access cloud-provisioned, SSO-enabled compute environments. Prepared interactive notebook-based Python and R material.
Automated and optimized scheduling Bar/Bar Mitzvah dates for ~130 students via mixed-integer linear programming.
Launched a carpool optimization tool at grouptherenow.com. Minimizes total drive-time sum across groups of 2-100. Configured for organizations.
Re-implemented LAPD’s “hotspot”-generation algorithm. Compared hotspots to historical arrest, citation, and crime report data from the City of Los Angeles. Contributed results the community-generated report "Predictive Policing in Los Angeles".
Developed a model for purchasing, production, shipping, and holding over a factory-warehouse-retail system. Forecasting using classical signal processing, regression, and machine learning. Implemented MVP in MATLAB.
Awarded "Llewellyn Bixby Mathematics Prize," 2014: to the student with highest achievement in the Mathematics department.
Optimizes decisions that affect compartment flow parameters in discrete-time SIRD disease progression model.
Extends strategies for algorithmic fairness from the machine learning community to a resource-allocation optimization setting.
Formalizes connections between the Boltzmann Machine Distribution and unsupervised learning based on sparse coding.
Explores unsupervised anomaly detection in video data using dictionary learning and sparse coding. An NSF-funded REU.
Awarded "Outstanding Presentation Award" by the Joint Mathematics Meeting, 2014: top 15% of undergraduate groups at JMM.