Skip to main content

Projects

Projects that required a method to be developed and tested. Some delivered analysis to support a decision; others became rules, workflows or tools used repeatedly.

01

Commercial decisions

THE CHALLENGE

Ordering depended on customer history, stock observed in the field, seasonality, pack sizes and commercial rules. Analytical support had to help the commercial decision, not replace it.

MY CONTRIBUTION

I developed the estimation logic that combined those inputs and built a Windows workflow around it, so proposals could be reviewed before they were sent on.

THE DELIVERABLE

A decision-support application used in live ordering cycles.

02

Forecasting and decision methods

03

Conversion and digital products

04

Research and evaluation

05

Data and analytical tools

Technologies and tools

Earlier work covered analytics, integration and automation. For new AIQU projects we choose tools based on the problem, data volume and intended use — not every platform listed is used on every engagement.

  • Earlier experience

    SQL, SQL Server, Oracle, C#, Windows apps, Excel/VBA — relational work, automation and API extraction on the cases above.

  • Data analysis

    Python, SQL, Polars, DuckDB, Parquet, APIs, dbt and data tests — shaping and checking datasets before a method is trusted.

  • Building tools

    TypeScript, Next.js, Supabase, Vercel, Git/GitHub and Docker — when the result must be an application or repeatable workflow.

  • AI integration

    LLM APIs, structured extraction, RAG and evaluations — when language or documents are part of the method and must be tested in use.

  • Processing at scale

    PySpark, Databricks or Fabric when volume needs distributed processing rather than a single-machine workflow.