Tom Kocik

Microsoft Certified Data Engineer

Data Engineer Consultant

paiqo · Austria Mar 2024 – Nov 2025
  • Fraud Detection solution for OMV Petrom using Azure Synapse, SQL Server, Logic Apps, Power Apps and Power BI.
  • HR support for TK Elevator using Azure Data Factory and Azure Databricks.
  • Cloud migration for Liebherr from on-premises SQL Server to a medallion architecture, star schema data model in Azure Databricks.

Data Scientist / Data Engineer

Australian Bureau of Statistics Apr 2022 – Sep 2023

Data Engineer

Australian Bureau of Statistics Feb 2018 – Jul 2018 · Nov 2018 – Jul 2021
  • Re-Engineering project to migrate Excel spreadsheets into the Macroeconomic Statistics Division's FAME ecosystem.
  • Internal System redevelopment from Visual Basic into FAME.
  • FAME Engineering support for the National Accounts Benchmarks Section.
Note: Space weather has been discontinued due to high operational costs. Notices were previously checked daily at 00:00 UTC using the Bureau of Meteorology's Space Weather API. Notices were specific to the Australian region and all dates and times are in UTC. Data was processed using medallion architecture in Azure Data Factory and Azure Databricks then loaded through a publicly accessible JSON file in Azure Storage.
Notices
Australian region daily A-index
Quality control (QC) is the ongoing process of monitoring a measurement system to make sure it stays accurate and consistent over time. Levey-Jennings charts are often used for QC to spot drift, trends or out-of-control points, often using Westgard rules to decide when a run should be flagged or rejected.

The Westgard Rules are:

1S3
One value beyond 3σ from the mean.
2S2
Two consecutive values either greater than, or less than, 2σ from the mean.
RS4
A difference between consecutive values greater than 4σ.
4S1
Four consecutive values greater than, or less than, 1σ from the mean.
10X
Ten consecutive values all greater than, or less than, the mean.

Below is an example of a Levey-Jennings chart with Westgard rule violations highlighted. The data was processed in Azure Databricks and displayed in Power BI. The AI interpretation was generated using a GPT-5 mini model deployed in Azure Foundry.

AI Interpretation
Lived
Visited
Bucket List