Miklos Magyar

Architected and built a full-stack secure backend for a global consumer product used by 600,000+ users, including iOS, iPadOS, and watchOS applications with ~99.9% stability and tens of thousands of daily users. Designed and scaled backend infrastructure from a single Firebase project to globally distributed, auto-scaling microservices on Google Cloud Platform. Implemented secure API gateway routing to a centralized REST API and forwarded requests to Dockerized microservices managed via load balancers and instance groups. Built asynchronous event-driven systems using Apache Kafka and later migrated to Google Cloud Pub/Sub for fully managed scaling. Integrated Elasticsearch for scalable search and implemented AI-enhanced features using OpenAI models for recognition, categorization, and insights generation. Shipped production LLM features with multilingual prompt and tool orchestration, deterministic runtime post-processing for brand voice enforcement, and automated behavioral evaluation

Delivered multiple client projects remotely, including: (1) Messenger Factory: white-label business messaging app with deep UI customization; (2) Medical management system for Austria (vaccines, medication and document tracking for doctors, hospitals, patients); (3) Powermum: German fitness app that dynamically assembled workout videos with synchronized audio; (4) Lego: interactive video pre-view prototype built on advanced geometric calculations.

Delivered multiple client projects through my own consulting company, including: a white-label business messaging app with deep UI customization (colors, icons, fonts); a medical management system for Austria covering vaccine/medication/document tracking across doctors, hospitals, and patients; a German fitness app that dynamically assembles workout videos with synchronized audio (adaptive sequencing, buffering, crossfade effects); and a 'Lego' interactive video preview prototype built on advanced geometric calculations.