OL Book a call

OGNJEN LAZIĆ

Senior Mobile & AI Engineer · Team Lead · AI Consultant

I build and ship production mobile, full-stack & AI systems — with products reaching 1M+ users.

View engineering work Book a consultation

PROOF

THREE PILLARS

01

ENGINEERING

5+ years shipping native Android, iOS & cross-platform apps to over a million users — leading mobile teams from first commit to production.

02

PRODUCTS

Original AI products, built end to end. UstavAI — the first AI legal assistant for Bosnian law, live with real users.

03

CONSULTING

Custom AI systems for businesses — RAG pipelines, autonomous agents, deep integrations into the tools you already run.

SELECTED WORK

LIVE PRODUCT · SOLO BUILD

UstavAI

Designed and deployed a full-stack AI legal assistant for Bosnian law — a production RAG pipeline with source-cited answers, sub-2-second responses and an in-browser PDF evidence viewer.

ustav.ai ↗ PRODUCT · IN DEVELOPMENT

Verba

On-device AI meeting notes for iPhone — recordings transcribed and summarised by models running entirely on the device, so nothing leaves the phone.

See the product → AI ASSISTANT · SHIPPED

Real-Time AI Assistant

Native Android AI assistant for a travel product, built with Kotlin, Jetpack Compose, GraphQL, WebSocket and AI-driven server-side rendering.

View related LinkedIn experience ↗

INSIDE USTAVAI

The first AI legal assistant for Bosnian law — designed, built and deployed solo, and now answering real questions in production.

The problem
Legislation sits in thousands of pages of scanned PDFs. Citizens and junior lawyers can ask a question in plain language, but nothing could answer it and point at the exact article it came from.
My role
Sole engineer and product owner — retrieval architecture, backend, frontend, deployment and the ongoing production operation.
Key decisions
Retrieval over fine-tuning, so every answer cites the statute it came from and can be checked. FAISS vector search with Cohere reranking to cut retrieval noise on near-identical legal language. Redis caching and rate limiting to hold responses under two seconds while keeping inference costs predictable.
Stack
FastAPIOpenAI embeddingsFAISSCohere rerankGoogle GeminiRedisRailwayVercel
Result
Live with real users. Source-cited answers in under two seconds, an in-browser PDF viewer with text highlighting and conversation history — and coverage on national television.

Try UstavAI live ↗ Watch the TV coverage ↗

EXPERIENCE

  1. 2025 — Present

    Full-Stack Engineer & Team Lead

    Hospitality & access-control platform · Dubai, remote

    Building a community access and control system with a cross-platform React Native app, and designing an AI multi-channel reservation system — owning the AI infrastructure and its end-to-end automation. Also delivering the short-stay rental platform with companion tvOS and Android TV apps.

  2. 2024 — 2026

    Senior Android Engineer & Mobile Lead

    Travel-tech startup · Dubai

    Led a team of mobile developers, owning planning, technical direction and feature delivery. Built a native tourist-guide app in Kotlin and Jetpack Compose with an AI assistant powered by GraphQL and WebSocket, plus AI-driven server-side rendering for dynamic content.

  3. 2023 — 2024

    Android Engineer

    US consumer sleep-tech app · 1M+ active users

    Maintained and extended a smart-bed application serving over a million active users in the US market — customer-facing settings and sleep-analytics features, and a Kotlin + Jetpack Compose modernisation of a legacy codebase.

  4. 2021 — 2023

    iOS & Android Developer

    TV & streaming

    Built a crypto wallet app from zero to production in native Kotlin, and an Android remote-controller app from scratch to first release with client–server communication to the TV set. Refactored an IPTV iOS app in Swift and UIKit for stability and maintainability.

  5. 2020 — Present

    Independent Contractor

    Freelance · mobile & AI

    Shipped an iOS IPTV app in Swift and SwiftUI with live streaming and TV archive, reaching 10K+ daily active users. Built native Android and iOS museum apps using image recognition for exhibit identification, and delivered a government-issued step-tracking app in Flutter on Firebase.

BSc Software Engineering — Faculty of Technical Sciences, University of Novi Sad · GPA 8.9/10 Download full CV ↓

HIRING, OR BUILDING
WITH AI?

Get in touch Download CV