Case Studies

Infrastructure projects framed by product impact.

These are the kinds of problems I enjoy solving: operational drag, release friction, and risky migrations.

Featured Project

GitLab AI MCP

GitLab AI MCP

A high-performance, containerized Model Context Protocol (MCP) server for integrating with self-hosted GitLab, featuring local AI triage and token-efficient data bundling.

-80%
LLM context reduction
100%
Local inference

Outcome

Privacy-first CI/CD debugging and 80% reduction in LLM context usage via intelligent response bundling.

GitLab MCP Python Docker AI
View on GitHub

Pipeline Throughput Upgrade

Pipeline Throughput Upgrade

Our release pipelines were getting slower as the team grew. I refined build caching and runner orchestration so developers stopped waiting around for feedback.

-45%
Build time
-60%
Queue wait

Outcome

Shorter feedback loops and fewer deployment bottlenecks.

GitLab BuildKit Terraform

Internal project — details available on request

Operational Reporting Automation

Operational Reporting Automation

I used serverless AWS primitives to automate recurring reports that were eating up hours of manual work every week. Less copy-paste, more time for actual engineering.

12/wk
Hours saved
-90%
Error rate

Outcome

Less manual reporting and more repeatable cloud operations.

Lambda EventBridge Python

Internal project — details available on request

Zero-Downtime Linux Migration

Zero-Downtime Linux Migration

Moving 120+ servers from CentOS to AlmaLinux without breaking payment-facing services was one of the trickiest projects I have led. Staged rollouts and service validation were our safety net.

0 min
Downtime
120+
Servers migrated

Outcome

Modernized estate without disrupting payment-facing services.

Linux Migration Compliance

Internal project — details available on request