Jovan Jevtić

After more than a decade of practical engagement with computers, and almost a decade of writing software, I can present my technical knowledge as the result of years of consistent dedication to every aspect of software development — from basic logic and hardware limitations to the complex architecture of distributed systems.

My starting point was a simple fascination with creating something from nothing using code; over time it became a professional orientation in which I mastered almost every segment of modern software development.

Already at twelve I built my first static web pages — a field of experimentation, but also an introduction to systematic, structured thinking. As my interest grew, focus shifted to dynamic systems, where I began developing my own frontend applications, paying special attention to interface, experience and accessibility.

Soon after, backend logic became the natural next step: building APIs, implementing business logic, and solving the problems that appear when applications must talk to databases, manage users, respect security and scale. What shaped me most was going through every phase independently — idea, design, implementation, testing, deployment, optimization — which forced me to understand how each layer works and how they coexist.

Over time I built solutions that resist simple patterns — web portals for news agencies, e-commerce with complex inventory and payment logic, internal analytics tools, and services drawing on many APIs: financial, crypto markets, health standards. Connecting that many sources and interfaces meant constantly minding performance, security and scalability.

In recent years I have focused intensively on mobile, with Expo and React Native at the core — not extended web interfaces but independent products using geolocation, offline capability, authentication, push notifications and the camera, with a deep grasp of the whole ecosystem, from state and local databases to UI performance and store distribution.

Authentication and security hold a special place: I have always built my own login systems rather than settle for ready-made libraries, designing flows for every scenario — multiple devices, automatic token renewal, abuse protection, identity verification — tested on real users in production.

My deepest competency is broader architecture — systems that grow and live for years without falling when users multiply tenfold. One result I am proud of: verifying authentic physical locations without the Google Places API (now largely withdrawn from open use), by combining web scraping, AI analysis, local heuristics and many public and closed sources.

A special focus is the practical application of AI. Within my startup I built an AI assistant that personalizes the experience through contextual content — connecting people with events, places and points of interest by behavior, location and interest — combining data analysis, semantic processing, geolocation and vector search. The work spans LLM integration through APIs and retrieval-augmented generation: embedding models, vector databases and proprietary filtering to search by meaning, not form.

I am comfortable building the interface logic that talks to LLMs — conversational flows, fallback mechanisms, relevance evaluation — for real control over AI behavior. And beyond practice: during my software-engineering studies I chose electives in machine learning, statistical modeling and AI for a deeper mathematical foundation.