A personal study tool in progress: an app for creating and drilling flashcards, built with Spring Boot and React. I'm building it to revise for the exams in my standalone courses — and to go deeper into modern, spec-driven development with Claude Code and OpenSpec. The foundation (monorepo, API and database) is in place; the rest is on a clear roadmap.

• Monorepo: Spring Boot backend (Java 21), React frontend (TypeScript/Vite) and PostgreSQL.
• Spaced repetition with the SM-2 algorithm for efficient review (planned).
• AI-generated cards from text and PDF via the Claude API (planned).
• Built step by step along a ten-step roadmap — the foundation (step 1) is done.

A thesis project in system integration that focuses on combining AI, news analysis, cryptocurrency and notification services into a single platform. The project is built on a modular microservice architecture in Java/Spring Boot and shows how data from several APIs and RSS feeds can be processed, summarised and distributed automatically through secure integrations.

• Integrates news feeds, AI analysis, crypto payments and notification services.
• Built with Spring Boot microservices, containerised and automated with CI/CD.
• Demonstrates how modern integration patterns can be turned into a working product.

A web application that helps teachers and students create and run quizzes in a modern and easy-to-use way. The project combines AI assistance with a clear structure so that users quickly find what they need.

• Question order can be sequential, reversed or random.
• On incorrect answers, a “Explain with AI” button appears with an AI-generated explanation.
• Incorrectly answered questions reappear later in the quiz after 10–15 questions for repetition.
• Text-to-quiz tool: paste text and generate quiz questions automatically, with language options and saving.
• OCR support: read text from photos or uploaded images (for example a book page) and use it to create new questions.

An e-commerce system where orders, products and customer data are handled by separate services but work together seamlessly. The goal is to show how a larger solution can be split into smaller services that are easier to maintain.

• Order flows use correlation IDs so each step can be traced across microservices.
• Error responses are returned as ProblemDetail to give clear feedback to clients and monitoring.
• JWT-based calls are validated in each service before data is exposed to users or other systems.
• Digital products are delivered automatically after purchase through an entitlements flow integrated with the user service.
• All services are orchestrated with Docker Compose so the entire system can be started and updated with a single command.