Jake Westfall Email & Phone Number

Austin, TX, US

US

US

Carpinteria, CA, US

• Contributed to the development of a feature store to facilitate the discovery, reuse, and quick retrieval of derived data by engineering teams throughout the company. Written in Java using Spring Boot, and using Flink.
• Built and launched a customer-facing service for summarized data reports. Developed a micro-frontend in Typescript using React, paired with a backend in Python using Flask and MySQL.
• Developed CI/CD pipelines to streamline Java jar and Docker deployments, leveraging Helm charts on Kubernetes.
• Maintained and expanded the capabilities of an internal data platform built on PHP with Symfony and a Neo4j graph database.
• Managed Airflow tasks for ETL processes, orchestrating Docker containers on a managed Kubernetes cluster, mainly involving database sources and Kafka sinks.
• Collaborated closely with data science teams to transition Python prototypes into production, addressing performance and stability issues.

Atlanta, Georgia, US

Atlanta, Georgia, US

• Designed and wrote Python simulations of our production web scraping process. These simulations were used to model how the production system would respond to various proposed changes
• Used Bayesian optimization to find the simulated scraping system configuration that would maximize both “efficiency” (performance under good conditions) as well as resilience to failures
• Using this simulation & optimization approach, I devised a new scrape prioritization algorithm that, when implemented in our production scraping system, hugely increased our data collection metrics. Crucially, this.
• Designed the database tables and ETL processes necessary to support commingling raw data samples with “inferred samples” derived from a predictive model
• Helped triage “fires” that arose in production; investigated which systems were affected and the extent of the damage, allowing engineers to focus on fixing the root issue

Atlanta, Georgia, US

• Developed demand forecasting models for both online and store products
• Developed pricing algorithms to maximize expected profit, subject to business constraints
• Designed and conducted randomized A/B tests of different pricing procedures
• Helped optimize bulk purchase discount amounts based on statistical models of bulk sales
• Worked on improving classification models that identify matching product webpages
• Co-developed and maintained an internal Python package for common data science needs

Austin, TX, US

• Co-created the open source Python package bambi, a facade/adapter for easily fitting Bayesian multilevel models using Stan or PyMC3 backends
• Published peer-reviewed journal articles on various statistical topics (see Publications).
• Served as statistical peer reviewer for papers in quantitative psychology
• Deployed analyses and simulations to Lonestar 5 supercomputer at University of Texas

Doctor Of Philosophy - Phd, Social Psychology

Bachelor Of Science - Bs, Psychology