Build
Create and test dependable data pipelines, models and quality checks that make raw information useful.
Data Engineering · SQL · Python · Snowflake · dbt | AI Search · GEO · AEO · SEO
I make complex data simple.
I build and validate reliable data pipelines with Python, SQL, Snowflake and dbt. Alongside data engineering, I investigate how search and AI systems find, cite and recommend information through SEO, AEO and GEO research. Both fields require reproducible methods and inspectable evidence.
Independent portfolio projects · Methods, evidence and limitations documented · Open to remote opportunities from Italy.
What I do
Investigate
Study how AI Search finds, cites and recommends information, including SEO, AEO and GEO questions.
Explain
Make the question, evidence and limitations clear to people who do not need to know the tools behind them.
Selected work
Projects built around evidence
E-Commerce Revenue & Operations Analytics
End-to-end analytics pipeline: 28 dbt models, 249 passing tests and a four-page Power BI report. Includes order lifecycle and financial reconciliation analysis.
Read the financial reconciliation case study →Inspect code & report ↗
Insurance Claims Analytics
SQL analysis of 10,000 insurance claims, exposure concentration and data quality, with an executive Tableau dashboard.
Read the insurance claims case study →Inspect project & dashboard ↗
SEO/AEO/GEO Optimization
An independent Notion case study moving from public baseline and query fan-out to diagnosis, proposed optimization and controlled measurement design.
Read case study →AI Search · Benchmark protocolAI Search Competitive Intelligence
A controlled benchmark across Notion, Asana, ClickUp and monday.com using 35 frozen prompts, first-party evidence and auditable scoring rules. AI response observations are pending.
Explore benchmark →Also exploring: How evidence shapes decisions about supplier access and data risk. Read the independent Cyber Risk & IAM case study →
How I think
Evidence first. Claims second.
I keep observed facts, analysis, proposed changes and measured outcomes separate. That matters in data work, and it matters just as much when evaluating visibility in AI-generated answers.