The systems behind
the experience.

The foundations behind a retail data platform.

Between 2012 and 2014, I co-developed the SQL infrastructure for a flagship retail data platform at a Circana predecessor.

The work brought together SQL Server, T-SQL and SSIS processing. My wider engineering experience involves millions of aggregated EPOS rows per retailer, per week, with close attention to performance and how the processing behaves.

My part

Co-development of the SQL infrastructure.

The focus

The processing foundations behind a longstanding retail data platform.

• SQL Server
• T-SQL
• SSIS
• EPOS processing
• Performance tuning

A working view of production.

I created an operational viewer and dashboards to bring together data receipt, retailer trends, quality checks, processing progress and environment information.

The overview includes recent weekly retailer and overall trends, with drilldown into retailer and category detail. Daily and weekly views help investigate where feeds are falling behind or where a result needs closer attention.

My part

Built the viewer, monitoring dashboards and drilldown investigation views.

The focus

Visibility across receipt, processing and delivery, with practical tools alongside the checks.

• SQL Server
• Trend analysis
• Data quality
• Latency monitoring
• Internal applications

Understand the change before it goes live.

I created an AI-assisted comparison tool to investigate the impact of proposed retailer data restatements before publication.

It compares live and test data, moving from the overall picture through different product and market views. The investigation can reach week, store and product detail to explain the drivers behind a difference.

My part

Built the comparison tooling and the investigation workflow, using AI-assisted development.

The focus

Explaining the impact of proposed data changes before publication.

• Live-versus-test comparison
• Restatement analysis
• Drilldown investigation
• AI-assisted development

From requests to self-service.

I built self-service job controls into the operational viewer, bringing existing SQL and Windows batch jobs into a place other teams can use.

Previously, colleagues would ask my team to run routine attribute-maintenance or extraction processes. They can now place the relevant input file and launch the job through the viewer themselves.

For the extraction jobs, the underlying Python scripts were written by a colleague. My contribution was the viewer and controls that make those processes available alongside the wider operational tools.

My part

Built the viewer and job-control workflow; integrated existing processing and scripts.

The focus

Other teams can run routine jobs without waiting for someone in my team to execute them.

• Self-service job controls
• SSIS workflows
• SQL Agent jobs
• Windows batch jobs
• Flask | AI-assisted

Seven meeting skills for an early Claude pilot.

Within a company Claude pilot, I developed all seven meeting skills created by our pilot group, contributing both the ideas and their implementation.

The pilot is exploring how reusable skills could become part of a wider company library. My contribution sits within the meeting-skills group.

My part

Developed all seven meeting skills created within the pilot group.

The focus

Reusable AI workflows for meetings.

• Claude pilot
• Reusable skills
• Meeting workflows
• Idea development

A different approach to product attribution.

My work has also included Naive Bayes product attribution, applying a classification technique to a real data problem.

It is another example of the part I enjoy: looking beyond the usual approach, understanding what a technique can offer, and bringing it back to the task at hand.

My part

Work on Naive Bayes product attribution.

The focus

Using classification in a product-data context.

• Naive Bayes
• Product attribution
• Classification