Case Study
Infobip: Agentic Product Engineering
Problem
At Infobip, I saw the gap between design instruction and production-ready UI as a systems problem. Product and design teams needed a faster path from natural-language intent to high-fidelity prototypes without adding translation loss, visual drift, or more design debt.
Process
I architected and deployed the first custom Cursor LLM-based tool at Infobip, designed to ingest our proprietary design language and turn plain-language prompts into brand-aligned prototypes. Alongside that, I used React, TypeScript, and agentic workflow patterns to accelerate discovery, prototyping, and implementation-quality handoff.
Solution + Outcome
The system enabled 100+ product-development teammates to generate high-fidelity prototypes from natural-language instructions while staying aligned with Infobip standards for reusability, accessibility, and UI quality. It also established a stronger foundation for agentic product engineering across discovery and delivery work.