Crafting digital experiences that matter.
Full-stack engineer who thinks in problems and builds until they're solved. I observe the atom to understand the universe, and the code to master the product.
Featured Work
MechanicAI
Vehicle diagnostics shouldn't require a mechanic visit to understand. MechanicAI takes a plain-language description of a car issue and returns a structured, progressive diagnosis through a 4-step AI pipeline — with subscription billing and full session history.
4-Step
AI Diagnostic Pipeline
2-Tier
Subscription Tiers
Drawgether
A real-time collaborative drawing platform built on Next.js 15 and Supabase Realtime. Features a custom auth system with TOTP, cursor-based infinite scroll, and live canvas sync via broadcast channels — with drawings auto-published on session end.
Real-Time
CANVAS SYNC
Custom
AUTH + TOTP
Recognition & rewards
About me
Luka
I'm a full-stack developer from Serbia. I build end-to-end, schema to UI, with a focus on real architecture.
Currently studying Computer Science at Metropolitan University. Outside the curriculum, I've shipped real-time platforms, AI diagnostic tools, and research that won Best Paper at IEEE STEC.
When I'm not building, I'm usually picking apart something that caught my eye and figuring out how it was made.
FAQ
Luka Rakic is a full-stack developer based in Serbia, building production web apps end-to-end — schema to UI — with Next.js, Ruby on Rails, and TypeScript. Recent work includes AI-powered SaaS products and realtime collaborative platforms.
Next.js and TypeScript on the frontend, Ruby on Rails or Supabase on the backend, PostgreSQL for data, and OpenAI's GPT models for AI-integrated features. Case studies below cover the specific decisions behind each stack choice.
Yes — open to both full-time roles and freelance projects. The fastest way to reach out is by email at hello@lukarakic.me.
Two recent examples: MechanicAI, a Rails API powering AI-driven car diagnostics with GPT-5 and Paddle billing, and Drawgether, a realtime multiplayer drawing app built with Next.js and Supabase Realtime. Full case studies are linked in the Work section above.
Yes — research work that won Best Paper at IEEE STEC, alongside ongoing Computer Science studies at Metropolitan University.