SQL Server engineer since 2008. I make databases fast and safe to change, and I run my work AI-first.
SQL Server
- sql-deploy-pipeline — build-once, deploy-per-environment GitHub Actions for SDK-style SSDT projects, with a per-action deploy ledger
- SSDTPowerTemplate — how I structure SSDT solutions: SDK-style projects, tSQLt, run-once scripts
- Benchmarks: 16 reproducible SQL Server experiments, each in a disposable SQL Server 2025 container, every result re-run on bare metal. Write-ups at dbm.lt/blog; harnesses available on request
- Load lab: a .NET 8 multi-connection SQL Server load driver with client- and server-side measurement and a live dashboard (private)
AI and infrastructure
- Homelab on Proxmox, defined in Terraform + Ansible, with an LLM failover proxy over local GPUs, RAG over its own docs and a kanban-to-agent dispatcher, operated through Claude Code agents. Architecture write-up coming on dbm.lt
- vs-code-snippet-generator — VS Code extension, 37k installs
Elsewhere: dbm.lt · Stack Overflow · LinkedIn

