My work
I work on enterprise platforms using Java and PostgreSQL. Here are a few cases from production, with the changes I worked on and how they turned out.
Java · Batch processing
A defect left more than 3,000 contracts incorrectly suspended. Their payment receipts were not being submitted for collection, putting €60,000 at risk.
I fixed the original Java batch process and added filtered execution to select and recover the affected contracts. We reused that recovery functionality in later incidents.
I also corrected the affected data and followed up on contracts and payments, working with managers and the joint venture.
Current work · PostgreSQL
Search and export were both struggling under load. After improving the queries, we began questioning whether large exports should run through the panel at all.
Read case studyOver 60s → about 4sSame search query and data in initial tests. Changes are still in testing.
Recurring incidents
In March 2026, I helped identify five recurring root causes with the L2 lead. I contributed to changes in the affected flows, supervised their introduction into production and helped address problems missed in testing.
Across the team’s work, the backlog fell from 284 to 175 between March and August 2026.
Three product launches
I monitored contract volumes and API performance using PostgreSQL, Grafana and logs during launches in September 2025, December 2025 and March 2026.
By the March launch, I was leading the daily data presentations to managers and Operations, explaining what we were seeing in production.
Since April 2026, I’ve led two developers in our production maintenance team. I assess incoming issues and assign fixes, staying involved when colleagues need technical support. I also manage stakeholder expectations and follow up on response and resolution deadlines.
Before my current role, I was a Software Engineer Intern at NTT DATA from October 2023 to June 2024, working on backend maintenance and debugging.
An earlier project
For my bachelor’s thesis, I built a local game and its interface in Python. It asked questions to narrow down the animal using Bayesian inference, entropy and information gain.