09/10/2026 updated


100 % available
Analytics Engineer | Power BI, SQL, Databricks
Val de Briey, France
Worldwide
LLMData AnalysisBusiness IntelligenceBusiness Intelligence DevelopmentGoogle BigQueryExtract Transform Load (ETL)Python (Programming Language)Microsoft Business IntelligenceOracle Business Intelligence Enterprise EditionMicrosoft Power BISupply Chain ManagementSQL DatabasesStatisticsDatabase EnginesSalesforce Tableau
I build the layer between raw operational data and the reports people actually use. Five years in Kroger Supply Chain Analytics writing the SQL, Python and PySpark behind operations at 35 distribution centers, after three years running continuous improvement at a Kroger dairy plant.
Most of what I shipped replaced something manual. Consultants had left an Alteryx dataflow for inbound service analysis, and it had two weeks before it was switched off. I rebuilt it as six PySpark notebooks on Databricks pulling from seven sources, including DB2, Azure Synapse, SQL Server and SharePoint. A three-day weekly process came down to 40 minutes.
Another was 75,000 scanned bills of lading in inconsistent formats, some of them handwritten. I fine-tuned an EfficientNet to find the regions, ran OCR, and used Gemini to normalize the addresses. That identified 2,800 freight lanes worth converting to FOB and has supported more than $1M in captured allowances.
On the Power BI side, I built the dataflow measuring off-guide transportation spend. The business maintains its classification logic in SharePoint, Databricks joins it to bid data, and the report shows what going off the routing guide actually cost. I later prepared that semantic model for Power BI Copilot with metadata, system instructions and a 30-question validation bank, and took it through AI governance approval. Other reports tracked inbound trucks against the 180-minute detention threshold and prioritized fleet warranty claims from IBM Maximo work orders.
At the dairy plant the work was statistical process control on production line data and a Bayesian competing-downtimes model to find which components were failing.
The tools behind all of that are Power BI (DAX, Power Query), SQL, Python, PySpark, Databricks, Azure Synapse, SQL Server, DB2, BigQuery, PostgreSQL, Tableau, Excel, Git, REST APIs, MLflow, OCR and LLMs. Industries so far are grocery retail, distribution and warehousing, transportation and fleet maintenance, and dairy manufacturing.
MS Statistics, MBA. Native English, French at B1. I'm based near Metz, work remote-first, and can come on-site around Metz for kickoffs and workshops. Available now.
Experience
Professional experience5 - 10 years
Freelancer since2026
Languages
EnglishNative speakerFrenchBasic knowledge