Hey there!
I’m a Staff Backend Engineer at GitLab, working on the infrastructure that runs GitLab CI/CD. Most of my work lives at the intersection of Ruby on Rails and PostgreSQL at scale. I focus on large-table partitioning, query performance, and schema evolution without downtime, while building the operational tooling required to support billions of rows.
I’m a database and backend maintainer in the GitLab codebase and a regular reviewer across the Verify section. A meaningful part of my week goes to mentoring, pairing, and writing documentation that other teams can adopt without needing to ask questions in Slack. I work in Ruby and Bash by default, reach for Go and Elixir when they fit, and have a soft spot for replacing any of them with a well-aimed shell one-liner.
Selected work
Pipeline partitioning. Authored the architectural blueprint and led the multi-year implementation that partitioned the largest CI tables (billions of rows) on GitLab.com without downtime. The work shipped as a sequence of incremental migrations and produced reusable database tooling for asynchronous foreign keys and constraint validation that other teams now rely on.
CI database size reduction. Identified that a large share of the CI database was duplicated metadata and designed a deduplication strategy projected to roughly halve its size. The same effort introduced pipeline retention, giving customers and GitLab.com itself a way to automatically clean up old pipelines.
CI queueing architecture. Contributed to a dedicated queueing table that decoupled job pickup from the main ci_builds hot path, then later led the migration of that system to work with partitioned tables and the Cells architecture. Both rollouts shipped without disruption to running pipelines.
parallel:matrix. Implemented support for defining a Cartesian product of variables on a single job, one of the more-requested keywords in .gitlab-ci.yml at the time. It is now part of how most non-trivial pipelines fan out test, build, and deploy work.
!reference. Added the !reference YAML tag to GitLab CI, letting users pull in configs from other parts of their configurations files without flattening or losing structure. It removed a long-standing rough edge in writing reusable pipeline templates.
Experience
I joined GitLab in August 2019 as a Backend Engineer in the Verify section, where I now work as Staff Backend Engineer. Before that I spent four years at Sparktech in Bucharest, the last as Technical Lead, building Ruby on Rails applications for content management, recruitment, and online learning, as well as an AI accelerator platform running on Kubernetes. Alongside the day job I taught a semester of Ruby on Rails at the Faculty of Mathematics and Computer Science of the University of Bucharest in 2018–2019, and earlier mentored students at DevAcademy, where I had started as a developer in 2013.
Education
Bachelor of Computer Science, Faculty of Mathematics and Computer Science, University of Bucharest, 2012–2015.