skip to content
kursataknc

$ cat about.mdx

About

Who I am

I approach software as rigorous architecture. My METU Mathematics foundation — algorithms, number theory, graph theory, optimization — left me with the habit of treating systems as architectures with provable boundaries rather than collections of files. That habit drives the platforms I build today: from leading the backend architecture at RevolArch to designing the deliberate, layer-driven lab hosted on this site.

End-to-end ownership is the default. I don't just write application code — I secure and maintain the infrastructure it runs on. As a freelance web developer and Linux sysadmin, I deliver full-stack production environments: modern web applications (Next.js, TypeScript, Tailwind) shipped concurrently with the deployment lifecycle — VPS provisioning, Nginx/SSL orchestration, Cloudflare network defense, and proactive host hardening with UFW and Fail2Ban.

Incoming: currently applying for the METU Cyber Security MSc (non-thesis track, starting September 2026). I treat security as a standard woven directly into the development loop—DevSecOps as the default, never a reactive patch. My focus is on integrating network traffic analysis, host hardening, and defense mechanisms natively into the codebase, backed by a solid mathematical foundation.

What's shipped

Each module I build is an isolated engineering slice, deployed with the intent of demonstrating scalable and auditable logic.

  • The Lab (/lab): a self-hosted multi-module web lab built into this site. Per-request CSP nonces, CSRF protection, consent-gated browser fingerprinting, SSE-streamed server telemetry, and a privacy-first weather pipeline. Built with strict pure-logic split + Zod-validated content, all running behind a hardened Caddy reverse proxy.

  • Infrastructure: a resilient containerised deployment topology managed via Infrastructure-as-Code principles, served behind Caddy + Let's Encrypt auto-TLS, deployed end-to-end through GitHub Actions.

Engineering principles

  • Privacy by Default

    Browser-side data collection is gated behind explicit consent. Progressive disclosure for invasive probes — Canvas / audio fingerprinting, WebRTC IP leak detection. The default state is zero collection.

  • Strict Validation & Layering

    Zod schemas guard every API boundary, environment file, and external response. Invalid data shapes fail loudly at build time, preventing state corruption from silent type drift.

  • Documented Engineering

    Every shipped lab module is paired with a documentation article explaining the architecture and the trade-offs that produced it. Decisions are auditable, not asserted as obvious.

  • Deterministic Testing

    Core logic — including the global terminal engine — is UI-agnostic and pure. This allows rigorous verification through Vitest unit testing and Playwright e2e suites running against production builds.

Stack

Backend & Architecture
Node.js · TypeScript · Python · Clean Architecture · DDD
Frontend
Next.js (App Router) · React · Tailwind · MDX
Infrastructure
Docker · Caddy · GitHub Actions · VPS · Cloudflare
Security
Per-request CSP nonces · CSRF double-submit · Rate limiting · PII redaction · STRIDE modeling
Tooling & Quality
Vitest · Playwright · Biome · Pino structured logs · Zod

Reach me

Email is the best route — see the contact page for direct details. Source and design specs live on GitHub.