Full Stack & Applied AI Engineer based in Belgrade, Serbia.
I build production web, mobile, data, and AI systems with TypeScript, Python, scalable backend services, and practical AI integrations.
I'm a software engineer with 6+ years of experience building web, mobile, backend, data, and AI-powered products.
- π’ Build production systems used across 100+ international service centers
- π€ Implemented AI-assisted CV analysis and candidate scoring for an internal recruiting platform
- π Built a React/Nest.js training platform used by 10k+ users
- π± Shipped Android apps with React Native, Expo, Kotlin, and Jetpack Compose
- π§ Integrated Gemini and DeepSeek into product features
- π§ͺ Built MediLense: a public drug-safety site with an LLM agent over 20M+ FDA reports, RAG, LoRA fine-tuning and 3D drug views
- π Open to remote roles and relocation
| Category | Technologies |
|---|---|
| Frontend | TypeScript React Next.js Angular Vite Tailwind CSS shadcn/ui Ant Design D3.js 3Dmol.js |
| Mobile | React Native Expo EAS Kotlin Jetpack Compose |
| Backend | Node.js Nest.js Express.js Python FastAPI Flask REST Socket.IO RabbitMQ |
| AI / ML | PyTorch Scikit-learn LightGBM Hugging Face LoRA RAG Embeddings Tool Calling LLM Evals MLflow Ollama Pandas NumPy Gemini API DeepSeek API |
| Database / Auth | PostgreSQL MySQL MariaDB SQLite Prisma Sequelize Supabase Clerk |
| Cloud / Infrastructure | AWS Cloudflare Workers AI R2 Docker Nginx CI/CD |
| Testing / Tools | Git Jest Vitest PostHog Bash Claude Cursor Codex |
| Payments | Stripe |
- π€ AI systems: CV analysis, candidate scoring, LLM integrations, structured outputs, RAG, and AI-assisted workflows
- π± Mobile apps: React Native, Expo, Kotlin, and Jetpack Compose apps with subscriptions, analytics, image processing, and Gemini/DeepSeek integrations
- π§Ύ Healthcare AI: MediLense, an LLM agent with tool calling, RAG over FDA drug labels, guardrails and evals over 20M+ FAERS reports, a LoRA-fine-tuned Qwen2.5-1.5B for drug name resolution, ranking models tracked in MLflow, and 3D views of each drug and its protein target
- πΈ Fintech & SaaS: Next.js, PostgreSQL, Stripe, financial workflows, and Python-based document processing
- π’ Enterprise systems: Production platforms used across 100+ international service centers and by 10k+ users
English C1 Β· Russian C2 Β· German, learning Β· Hindi, learning




