Business Analyst · Industrial performance · Data products
I build practical data and automation projects at the intersection of industrial operations, controlling and trustworthy AI.
My focus is not generic dashboards or autonomous agents. I am interested in systems that make operational decisions more traceable, measurable and reviewable.
- industrial performance and Actual / Budget / Forecast analysis
- Python, SQL, data quality and reproducible pipelines
- process improvement, KPI reporting and operations analytics
- practical ML and LLM workflows with strong human oversight
- European industrial, aerospace and logistics use cases — using only independent, public or synthetic data
| Project | What it demonstrates |
|---|---|
| OpsPilot AI | Evidence-first industrial performance MVP: Excel ingestion, deterministic quality gates, variance analysis, source lineage, Streamlit UI and automated tests. |
| KPI Report Generator | Validated CSV-to-Markdown KPI reporting with deterministic financial calculations and safe CLI error handling. |
| ETL Pipeline Automation | Deterministic synthetic CSV ETL with validation, rejected-record reporting, audit summaries and protected output paths. |
| Margin Analytics Lab | Synthetic-only category margin reporting with deterministic validation, ranked Markdown output and explicit accounting limitations. |
| Maintenance Work Order API | Local Flask and SQLite API for fictional industrial work orders, controlled status transitions and automated tests. |
| Turbofan Health ML | Synthetic-first Remaining Useful Life baseline with engine-level holdout evaluation and a C-MAPSS-compatible parser. |
| Predictive Maintenance ML | Synthetic machine-condition anomaly-detection baseline with validated parameters and transparent alert evidence. |
| Social Media Mental Health NLP Study | Synthetic-only text-classification research prototype with explicit non-clinical and non-diagnostic limits. |
| Logistic Challenge | Fictional transportation optimisation model with capacity, demand and feasibility tests. |
| Sales Dashboard Flask | Local-only Flask dashboard using embedded synthetic sales data and a tested health endpoint. |
I build a small set of complete, well-documented projects rather than inflating activity with artificial commits. The public projects are deliberately scoped as learning or portfolio work, with explicit limitations and synthetic/public data boundaries.
- Industrial performance and data quality — OpsPilot AI, KPI reporting and ETL validation.
- Operations workflows — work-order, cost-tracking and logistics optimisation exercises.
- Responsible industrial ML — synthetic-first RUL and anomaly-detection experiments.
- Safe local tools — explicit data boundaries, tests, source control and CI before public presentation.
- No employer, customer or confidential data in public repositories.
- No hardcoded credentials, tokens or private keys.
- Tests, reproducibility and limitations matter as much as features.
- Public code is a learning and portfolio record — not a claim of production readiness.
Python · SQL · Power BI · Excel · SQLite · Pandas · Streamlit · Flask · Git · Docker
Based in Hamburg, Germany. Open to industrial data, transformation and operations challenges.