I am
Architecting & building systems that think, move fast and perform miracles 🏎️💨🤖
I take products from a blank directory to production - full-stack architecture, AI agents and orchestration, and design that makes software feel alive.
Nuxt.js
TypeScript
Laravel
Python
React.js
HTML
Tailwind CSS
Vuetify
MySQL
MongoDB
GraphQL
Solana
AWS
Cloudflare
Playwright
ElizaOS
Codex
Prompt Engineering
Figma
Jira
Vue.js
JavaScript
PHP
Node.js
SCSS
Bootstrap
Three.js
PostgreSQL
Redis
Web 3
Docker
Apache
Vercel
GitHub Actions
n8n
Claude
AI Automation
SEO
Notion
Trello
100% Job Success Score
PLUS
FINKI Skopje
I architect software end to end, from a blank directory to a production system serving clients worldwide. For 10+ years I've designed and shipped full-stack platforms with Vue/Nuxt and Laravel: e-commerce, inventory systems, marketplaces, Web3 apps, and everything in between. Today I build systems that think - AI agents, orchestration pipelines, and automations that turn ideas into products at speed. As an AI Creator & Prompt Engineer, I pair engineering rigor with a creative flair: 3D experiences, generated media, and a keen design eye that makes products feel alive. If you can describe it, I can build it. Let's build something that feels like the future.
University 'Sts. Cyril and Methodius', Faculty of Computer Science and Engineering (FINKI), Skopje
Computer Science and Engineering
Undergraduate Studies Study program: 'Computer Science and Engineering - KNI'
Semos Education
Web Development
20486B: Developing ASP.NET MVC 4 Web Applications, 20483B: Programming in C#, 20480B: Programming in HTML5 with JavaScript and CSS3
Freelance
Vortex
FFInventory
ZapInventory
Searchiest
Netcetera
Professional services I offer to help bring your ideas to life
Full e-commerce store with admin CRM: product catalogue, cart, payments (card + crypto), inventory, order management, customer CRM, and analytics.
Custom website from Figma or brief: responsive design, SEO-ready, fast performance, domain/SSL setup, and deployment to any platform.
Custom Claude AI agent setup: MCP server, tool integrations (Trello, Jira, Gmail, Telegram), autonomous workflows, and autopilot pipelines.
7-step SEO audit: technical crawl, Core Web Vitals, keyword research, competitor analysis, on-page review, backlink audit, and prioritized action report.
Complete MVP development: database design, API, frontend, SEO, QA with proof per task, deployment, and full documentation handoff.
Delivered work and creative projects for clients
Trusted by clients worldwide - real stories presented within a rotating global view
Special discounts and offers from services I personally use and recommend
I'm currently open to new opportunities. Feel free to reach out if you'd like to collaborate!
Thoughts on software development, Vue.js, Nuxt, and building modern web applications.
From MVP to ProductYour AI agent works in staging but fails in production. Non-deterministic responses and cascading errors cost you users and revenue. There's a testing framework that catches these failures before they hit your customers — and it's not in any documentation.
From MVP to ProductMost developers still treat AI as a text-only autocomplete. That's leaving money on the table — and your users expect more. Here's the exact architecture to combine Vercel AI SDK 7 agents with a FastAPI backend for text, images, and audio in one streaming pipeline.
AI & Economic ImpactYour autonomous AI agents are making decisions right now that you can't trace. Under the EU AI Act, that's not just sloppy engineering—it's a liability that could cost 7% of global revenue. There's a structural pattern that solves this without slowing your agents down.
From MVP to ProductYour AI agent crashes silently, costs explode, and users see blank screens. Three retries won't fix it. There's a layered survival system that stops failures before they cascade — and it's not in any SDK docs.
From MVP to ProductYour AI agent forgets everything after 5 minutes of conversation. That's not just annoying — it's costing you revenue, trust, and the ability to scale. But there's a tiered memory architecture that makes agents remember context, past decisions, and user preferences across sessions. And it's simpler than you think.
From MVP to ProductYou've built the demo, but your AI app falls apart under real traffic. Slow responses, spiraling costs, and brittle agent loops are killing your product. There's a proven six-layer architecture that fixes all three—and most teams skip layer four entirely.
From MVP to ProductYou're manually wiring design tokens into every component, and it's costing you hours per sprint—while the UI still drifts from specs. There's a smarter way: runtime-assembled interfaces that pull design systems live, adapt to user intent, and ship faster. Here's how to build them without the complexity.
From MVP to ProductYou've spent months shipping features, but now AI promises to automate workflows—yet every tutorial leaves you with hallucinating endpoints and broken state. There's a pattern that keeps agents predictable, auditable, and safe. And it doesn't require rewriting your entire stack.
AI Tools for Small BusinessYour design system is drifting. Every AI-generated component introduces subtle visual inconsistencies that snowball into a fractured UI. There's a new machine-readable contract approach that locks in brand rules before a single line of code is written—and it's not in any tutorial.
From MVP to ProductYou've built a solid full-stack app, but adding AI agents often breaks everything—slow APIs, drift nightmares, and security holes you didn't see coming. There's a battle-tested approach that top teams use to ship agents without the chaos. And it starts with one critical decision most tutorials get wrong.
AI & Economic ImpactYour users expect instant AI responses, but JavaScript's single-threaded nature turns inference into a laggy, CPU-hogging nightmare. Offloading to the cloud adds 200ms of network latency per request. There's a way to run complex models directly in the browser at near-native speed—and it doesn't require rewriting your entire stack.
From MVP to ProductMost devs spend weeks wiring LLMs into their UI. You're stitching together streaming, state sync, and component rendering by hand. There's a faster way that cuts your build time by 80% — and it's production-ready today.
From MVP to ProductYour full-stack app is leaking sensitive data through every AI request. Without tenant isolation and audit logs, one prompt can expose your entire user base. Here's the gateway pattern that blocks leaks before they happen.
UncategorizedYou push code, merge it, and three hours later a bug surfaces in production. That feedback loop costs you time, sleep, and credibility. There's a smarter way: a real-time AI agent that reviews every PR before it lands — and you can build it in one afternoon.
UncategorizedYou're paying for servers that sit idle 80% of the time, yet your AI pipeline still crumbles under sudden traffic spikes. There's a smarter way to scale — and it won't drain your budget or your sleep.
UncategorizedYour notes are scattered across devices, apps, and folders. Every time you switch machines, you lose context and momentum. There's a way to build a personal AI knowledge base that syncs seamlessly and learns from everything you throw at it — no cloud subscription required.
UncategorizedYou've got the code, the database, and the API endpoints — but your app still feels dumb. Every manual task, every slow decision, every bottleneck is a leak in your pipeline. There's a way to wire multiple AI agents together so they coordinate your entire stack autonomously. And it's not science fiction — it's running in production right now.
UncategorizedYou're paying for idle servers. Your AI pipeline can't handle traffic spikes without crashing or costing a fortune. There's a serverless pattern that handles 1000 requests per second on a shoestring budget, and most developers are still building it wrong.
UncategorizedManually reviewing every toxic comment or image is killing your team's productivity and letting harmful content slip through. One flagged post can tank your brand reputation overnight. There's a faster, smarter way using AI pipelines — and you can set it up before lunch.
From MVP to ProductYou've built a stunning AI-Next.js app, but every deploy feels like Russian roulette—streaming responses fail, model outputs shift, and your tests break for no reason. The real cost isn't just bugs; it's lost trust and sleepless nights. There's a testing pattern that fixes this mess, and it's not in any tutorial.
