Andrew P. S. Email & Phone Number

Kirkland, Wa, Us

Kirkland, Wa, Us

Menlo Park, Ca, Us

* Project lead to improve "freshness" of data to ML models, saving $500M/year in 2023. Mapped all 20,000+ critical paths for data flows to ML training, and drove effort to improve reliability and performance of these data flows.* Migrated 2.5 million cpu-days and 440 PB worth of daily data warehouse workload from Hive from Data Warehouse to Spark and/or Presto. Saved $20mil/year in efficiency gains.* Ascended to Efficiency Ninja, helping manage a program of data efficiency across the company. Held regular office hours unblocking almost 100 teammate’s projects helping find ways out of stifling performance issues.* Helped automate the detection of, and helped automatically or assist in fixing thousands of inefficient pipelines, saving $53mil/year in 2021* Trusted to run a long-term project focused on operating and improving the single-most heavily used table in Meta's data warehouse, a privacy-related user mapping table critical for all user queries. Improved its landing time by two hours a day, saving 244000 pipeline-hours per day across the entire warehouse This led to improvements in job orchestration, saving $37 million/year in resource savings from fewer failed jobs across the whole warehouse.

Fremont, California, Us

Improve the execution of Hive data pipelines moving trillions of records, petabytes of data, and migrate them to run using more modern engines like Spark or Presto. In the first two months, saved almost 2000 CPU-days/day equating to ~ $1.5 million saved per year, and directed pipeline owners where possible on how to self-migrate, saving even more. Applied data science skills to drive work where it would have the most impact, collecting pipeline metadata, creating scoring model, classifying ones which would likely benefit from manual intervention to improve efficiency.

Advancing oversight capability of Amazon Pets group through the use of advanced analytics, data analysis, predictive modelling to optimize various supply chain costs, promote in-stock metrics, and understanding parity with competitors.

Kirkland, Washington, Us

Classified marketing leads for Microsoft’s C+E division to find ways to increase Known Leads, MAQLs, and SQOs.Reported on trends with Microsoft’s marketing lead funnel metrics from Marketo

Seattle, Wa, Us

 Developed risk reduction models to predict credit default potentials among Brooks customers. Used Python, scikit-learn, logistic regression.  Doubled loading speed for reports and dashboards through code improvements. Reduced EDW loading time by 90% by rewriting SQL stored procedure transforms for data quality/efficiency. Established unit testing of transact-sql (t-sql) transforms, coding standards, two-week Agile sprints. Reduced report generation costs by hundreds of man-hours through data visualization and advanced analytics of sales trends using Tableau. Worked with Sales Rep Managers and VP of Sales to address latency, slowness, bad data and other concerns. Developed Tableau Sales Dashboards fed by data warehouse data, live source data, and data fed from retailers. Created predictive model from thousands of runners' biometric data for a shoe recommender application. Model was built in Python using multinomial logistic regression and K-means clustering. Advised lab scientists on experimental designs, data collection, feature generation. Served on Enterprise Data and Advanced Analytics Committee promoting company-wide data-driven decision making.

Seattle, Wa, Us

 Led team of six developers to build clinical data warehouse upon IBM Netezza. Loaded, transformed billions of records, hundreds of tables, transforms, for clinical, financial, patient safety, and research projects. Doubled daily EDW refresh speed by rewriting SQL, reorganizing data structures, job steps. Coordinated with upstream teams from Epic, Cerner to save hours each morning on data refresh. Doubled productivity with developers by encouraging the use of test-driven development and trained others my data quality framework written in Bash and SQL. Standardized monthly release process using Bash scripting. Automated weekly releases. Improved self-service across the hospital by teaching SQL query and Tableau classes. Mentored analysts/developers on the platform. Designed and built Physician Profile end-to-end data pipeline for U.S. World and News Report. Create Hospital/Professional Cost data pipelines to support changes related to Affordable Care Act.

Seattle, Wa, Us

 Drove 100x increase in real-time EDW load performance of discrete clinical events using C#, micro-batching to SQL Server.  Architected financial, clinical data marts in SQL, SSAS, SSIS, including one for the Harborview Budget Director and the Daily Flash Report using SSRS and later Tableau

Tempe, Az, Us

 Created an estimated savings of $200K per year by improving nursing student admission processes and increasing speed of admissions by writing a web-based system to automate the complex processing of incoming Nursing students, using Java, JSP, Spring, Maven, Oracle, calling on various university and external web-APIs for coursework, grades, immunizations, exam scores. Played a critical role in implementing GE Centricity EMR for health clinics, transferring all patient data, and rewriting all extracts to provide data to grant providers including Title X. Automated time-consuming data transformation tasks using Transact-SQL previously performed by clerical staff. Optimized databases running on SQL Server, MySQL

 Helped design and build Head2Head.com, a fantasy football site for TimeWarner. Gathered technical specifications. Developed conceptual/logical/physical data models. Worked closely with graphics designers and implemented use cases. Followed Agile project management on all projects and led scrum.  Designed and optimized databases on MS SQL Server, MySQL, and Oracle. Used C#, Java, Transact-SQL, SSIS, Perl, and PL/SQL for data transformation.

Tempe, Az, Us

Reduced form handling errors by 90% by designing and building a complex student form tracking system in C# for ASP.NET to handle many time-consuming tasks within the College of Liberal Arts and Sciences.

Armonk, New York, Ny, Us

Developed SQL Server monitoring, anomaly & fault detection software for IBM clients. Written in C# and ASP.NET using a SQL Server database, the system monitored 3000 databases on 200 SQL Servers to managed data integrity for Honeywell, Moody’s Investor Services and American Express.

Master Of Science (Ms), Data Science

Professional Cerificate In Data Science, Data Science

Bachelor Of Science (Bs), Computer Science