Business analyst. Most of my work sits where requirements, data and testing meet: turning what stakeholders say into requirements someone can test, checking the data behind them with SQL and Excel, and seeing changes through UAT.
Before this I was a business analyst at Banking Circle (data validation, traceability, UAT planning), worked in customer care at Elevance Health, and did business systems analysis on an employee portal and an inventory system at Saklax Manufacturing.
These projects share one fictional client, Northstar Systems, a mid-sized SaaS company. The company and its people are made up. The code, the checks, the tests and the results are real, and each repository can be re-run from a clean clone.
Short version: three one-page case studies (PDF), about two minutes each.
| Project | What it covers | Tools |
|---|---|---|
| RequirementsIQ | Requirements for Northstar's product team. Stakeholder notes go in; reviewed requirements, user stories with acceptance criteria and UAT test cases come out, with approvals, traceability and a Jira import file. It includes an AI drafting step, and nothing counts until a named reviewer approves it. | Python, SQLite, Streamlit, Power BI, Jira |
| Billing migration reconciliation | Northstar moves billing to a subscription platform. Data profiling, source-to-target mapping, data quality rules and an independent SQL reconciliation across two mock cutovers, with a defect log, a cutover runbook and a go/no-go recommendation for the Controller. | SQL, Python, Excel |
| Support ticket triage evaluation | Should a language model triage Northstar's support tickets? Today's keyword rules, an in-house text model and two Claude models are scored on the same 500 held-out tickets against acceptance criteria committed before the first model call. Includes the labeling guide, every miss explained, a cautious business case and a gated pilot plan. | Python, scikit-learn, Anthropic API, pytest |