Reporting, pipelines and data models that run without manual rework.
I translate between business systems and the people who make decisions with them.
Eight years in analytics, five of them in enterprise operations. M.S. in data science, an actuarial and economics foundation, and ongoing work in applied AI.
I’ve inherited enough undocumented work to prefer solutions that document themselves. An ETL script you can read top to bottom is documentation. A hundred cross-referencing custom calculations is a problem you hand to the next person.
Forking a report to satisfy a local requirement is cheap once and expensive forever. Configuration and conditional logic inside a single product instead.
Refresh timestamps, row-volume patterns, validation checks. If a stakeholder notices first, the system was incomplete.
No access to the source I wanted. No license for the easiest tool. No upstream change on my timeline. Each one forced a design I wouldn’t have reached otherwise.
Hundreds of software partners, tracked in spreadsheets and rolled up by hand. I built and ran the platform that replaced it — reporting, business logic, ETL and entitlements — serving partners and internal leadership each at their own scope, for five years.
Tableau · Alteryx · row-level security · Salesforce / SAP / Oracle sources
Customer identifiers disagreed across enterprise systems, and automated reassignment kept moving a revenue metric leadership relied on. I traced the cause, made the case for a cross-system mapping, and built the reporting that pinpointed affected accounts so they could be resolved in batches.
A legacy Excel process nobody could fully explain. IT moved the platform data into the data lake; I reverse-engineered the existing reporting and built the ETL in Tableau Prep, combining the sources into what became the learning and development team’s primary data source for five years — and the foundation for every L&D dashboard built on it since.
Tableau Prep · Tableau · row-level security
A lending technology company had no BI environment. I stood up their first Tableau environment and embedded the reporting directly into Salesforce, teaching myself the HTML needed to make the embed work.
Tableau · embedded analytics in Salesforce · HTML, self-taught
A national retailer, 55,799 SKUs, three years of forecast demand, six candidate distribution centers. Greedy lowest-cost assignment filled 92.6% of demand by volume while leaving 98.6% of customers with an incomplete order. The fix isn’t a better solver, it’s a different objective.
Python · pandas · network design · capacity-constrained allocation
SAS data pulls linking complex claim logic to ICD codes and claim dates, feeding automated claims, audit and compliance reporting.
Consolidating Whoop and Strava into one training plan for the Detroit Free Press Half Marathon.
Pulling apart my own watch data to see what’s in it. Early stages.
Built by hand rather than on a platform — DNS, hosting and mail routing included.
I build automated data systems. Open to full-time roles and project work.
Or write directly to contact@ski.systems
ahpawlikowski@gmail.com