Software developer · Python, Java, SQL · Geneva
I build backends and data pipelines for domains where mistakes are expensive: personal data, payroll, legal deadlines, regulated records. Two years of hands-on development, on top of 12 years inside international organisations, where I worked daily with structured XML data, cross-system reconciliation, access management and incident resolution.
That combination is what I bring to a team: I read a regulation or a business process, turn it into explicit rules, and ship them as tested, reviewed, documented code.
📍 Geneva · Swiss B permit · French, English, Russian
patent-xml-validator - XML data validation
Checks PCT patent bibliographic data: schema validation with line numbers, 18 PCT business
rules (priority periods, numbering, kind codes, receiving office competence), XXE-safe parsing,
results stored in SQL and an HTML report for the people correcting the data.
Java 21 XML / XSD JDBC H2 JUnit 5 · CI with a generated report
headcount-data-quality - data pipeline
Consolidates HR and payroll exports from three systems, runs 15 data quality rules
(including AVS check-digit validation and HR ↔ payroll cross-checks), computes headcount,
FTE and staff movements in SQL and generates an Excel report for the statistics team.
Python pandas SQL openpyxl · 65 tests · CI on Python 3.11–3.13 · report built as a CI artifact
helpdesk-api - REST API
Ticketing backend with SLA due dates per priority, an enforced status workflow (invalid
transitions return 409), comments and a statistics endpoint.
FastAPI PostgreSQL SQLAlchemy 2.0 Alembic Docker · 30 tests run against PostgreSQL in CI
textstats-go - CLI tool
Text statistics from the command line with JSON output for scripting.
Go · CI on Linux, macOS and Windows · cross-compiled release binaries
Kekstagram · demo - front-end app
Image-sharing app in vanilla JavaScript with an LLM API for hashtag suggestions,
13 modules, input validation and XSS protection.
JavaScript ES6+ REST Vite
- Business rules isolated from transport and storage, each with its own tests
- Feature branches and pull requests; nothing merges without green CI
- Database schema versioned with migrations; drift checked in CI (
alembic check) - Validation at every boundary: API payloads, file imports, user input
- Reproducible environments with Docker; secrets kept out of the code
- Documentation that explains decisions, not only commands
Backend & data: Python, FastAPI, Pydantic, SQLAlchemy, pandas, SQL (PostgreSQL, SQLite)
Front-end: JavaScript (ES6+), HTML, CSS / Sass, Vite
Tooling: pytest, Ruff, Docker, Git, GitHub Actions, Linux · Go for CLI tools

