Third-year CS student who builds things to understand how they work. Backend systems, full-stack products, the occasional low-level rabbit hole. Just wrapped up a production internship, learning Go properly, and getting pulled into AI tooling whether I planned to or not.
Built during exam season, which is probably the worst time to build a scheduler — and also exactly why I did. It's a Google Calendar MCP server written in Go: add, list, delete events straight from any MCP-compatible client. OAuth2 auth, refresh token persistence, typed tool definitions. No vibe coding — I actually read the MCP Go SDK. Released three versions already.
Wanted to understand how real-time systems work, so I built a blogging platform around it. RabbitMQ for async notifications, Cloudinary for media, JWT auth — the kind of project where you keep pulling one thread and end up building a lot more than you planned.
Got curious about how video streaming actually works. Built an LMS from scratch — HLS adaptive streaming, Redis caching, proper auth. The interesting part was making it hold up under concurrent load without things quietly breaking.
My first real Go project. Wraps FFmpeg to pull frames out of video files from the command line. Simple enough to finish, complex enough to teach me something. Now adding goroutines for batch processing — really just an excuse to learn Go's concurrency model properly.