Backend Developer building secure, scalable software with Java, Spring Boot, and cloud-native architecture. Currently finishing my backend developer education while running two businesses on the side - so I care a lot about shipping things that actually work in production.

Gothenburg, Sweden · Open to backend and full-stack opportunities

• Eloize Audit Platform - Production system for Eloize AB that powers the AI Growth Audit product.
• Nordic Dev Mentor - A Spring Boot middleware proxying OpenRouter chat with four AI mentor personalities, paired with a Next.js frontend (Spring Boot 4, Java 21, Next.js 16, React 19, Tailwind 4, Railway)
• Vet1177 - A veterinary case management system with ABAC security (Java 24, Spring Boot 4, PostgreSQL, MinIO)
• Collaborating with teams using feature-branch workflows, PR reviews, and CI/CD pipelines

Backend: Java 21/24 · Spring Boot 4 · Spring Security · JPA/Hibernate · JWT (ES256) · OAuth2 Resource Server · REST APIs
AI & Integrations: Anthropic Claude API · Perplexity API · MCP servers · OkHttp · Spring WebFlux WebClient · PDFBox · Apache POI
Automation: n8n · Make.com · GitHub Actions
Databases: PostgreSQL (incl. JSONB) · Supabase · Azure SQL · MySQL · HikariCP
Storage & Files: Supabase Storage · MinIO (S3-compatible) · AWS SDK v2
Frontend: React · TypeScript · Vite · Tailwind CSS · Recharts
DevOps: Docker · Docker Compose · Railway · Vercel · Azure App Service · Testcontainers
Practices: REST API design · Async orchestration · ABAC authorization · TDD · SOLID · Clean architecture

Production system for Eloize AB that powers the AI Growth Audit product. A Spring Boot backend orchestrates Claude to generate multi-section audit reports from uploaded documents and client interviews, with two separate React frontends - one internal tool for consultants, one client-facing portal. Deployed and running in production on Railway + Vercel.

This is a commercial project - all repositories are private. Architecture and technical details are documented below. Happy to walk through the code and demo the live system in an interview.

Backend: Java 21 · Spring Boot 4 · Spring Data JPA · Supabase PostgreSQL · Anthropic Claude API · OkHttp · WebFlux WebClient · Jackson (JSONB mapping) · Railway
Frontends: React · TypeScript · Tailwind CSS · Recharts · Axios · Vercel
Storage: Supabase Storage for PDF/DOCX/TXT uploads (up to 25 MB) · PDFBox & Apache POI for server-side text extraction

Architecture highlights:

• Async AI orchestration - Dedicated newFixedThreadPool(3) for Claude API calls to avoid starving Tomcat threads. Generates six report sections sequentially with dependency passing (e.g. WORKFLOW_INVENTORY numbers feed into EXECUTIVE_SUMMARY)
• Graceful degradation - Section failures are logged and the pipeline continues rather than failing the whole audit
• Single-source-of-truth dashboard - DashboardService aggregates AI outputs into one endpoint so the UI doesn't recompute across sections
• Auto-generated workflows on publish - Publishing an audit synchronously triggers generation of 3–5 automation workflows, including n8n-compatible JSON templates
• JSONB domain mapping - Client profiles and audit metadata stored as PostgreSQL JSONB, mapped via Jackson with @JdbcTypeCode(SqlTypes.JSON)
• Public + authenticated endpoints - Clients access published audits via short access codes (ABC-1234); pre-interview forms use one-time tokens
• Manual schema management with ddl-auto=validate and explicit cascade delete logic in AuditService to match business semantics
• Slack bot integration - Protected API endpoints allow an internal AI agent to manage audits directly from Slack

Three deployment units:

• Backend (Java/Spring Boot) - REST API, AI orchestration, data layer → Railway
• Internal Portal (React) - consultant dashboard with file uploads, AI section review, and publication workflow → Vercel
• Client Portal (React) - client-facing app with access-code sign-in and Recharts visualizations → Vercel

API surface: 40+ REST endpoints across /api/clients, /api/audits, /api/audits/{id}/workflows, /api/files, /api/forms, and /api/public/* - full API documentation available on request.

A Spring Boot middleware that proxies chat requests to OpenRouter with four distinct AI mentor personalities, paired with a Next.js frontend in editorial Nordic design. Deployed as two Railway services with the backend reachable only via internal DNS - the API key never leaves the server.

Stack: Spring Boot 4 · Java 21 · Reactor WebClient · Next.js 16 · React 19 · Tailwind 4 · TypeScript

Highlights

• Hexagonal architecture - domain layer has no Spring imports
• Per-personality system prompts and sampling temperature
• Exponential-backoff retry with Idempotency-Key for OpenRouter resilience
• localStorage-backed conversation history with click-to-resume
• 14 backend tests (JUnit 5 + WireMock) and 26 frontend tests (Vitest + RTL)

Live demo

A full-stack web application for managing veterinary case records. Pet owners create and track medical cases for their animals, veterinarians manage and respond to cases at their clinic, and admins have full system oversight. Built as a group project with file attachments, real-time activity log, role-based access control, and a full REST API — all containerized and ready to run with a single command.

Backend: Java 24 · Spring Boot 4.0.4 · Spring Security · OAuth2 Resource Server · PostgreSQL 15 · MinIO (S3-compatible) · AWS SDK v2 · Docker Compose · GitHub Actions CI/CD Frontend: React 19 · Vite 8 · Tailwind CSS · Axios · jwt-decode Highlights:

• Policy-based ABAC authorization layer with per-entity policies
• Custom exceptions mapped to HTTP semantics (400/401/403/404/422) via @RestControllerAdvice
• Activity log / audit trail on all case events
• 150+ feature branches, 115+ PRs merged across a 4-person team
• Documented API.md and environment-driven configuration

Built with @lindaeskilsson, @TatjanaTrajkovic, and @johanbriger.

A cloud-native Applicant Tracking System demonstrating the full lifecycle of modern software development — from secure backend logic to automated deployment and real-time candidate management.

Stack: Java · Spring Boot · Supabase (PostgreSQL + JWT ES256) · React · TypeScript · Tailwind CSS · Railway · Vercel Highlights:

• Token-based security with Supabase JWT authentication (ES256)
• Row Level Security on the database layer
• Drag-and-drop recruitment pipeline with stage management
• Detailed candidate scorecards and role-based admin controls
• Fully automated deployment via Railway CLI and Vercel

Backend repo (mini-ATS) · Frontend repo (talentflow-pro) · Live demo

A full-stack meeting management application with a dark glassmorphism UI, Kanban board with drag-and-drop, dashboard with statistics, and calendar view.

Stack: Java 21 · Spring Boot · Spring Data JPA · Thymeleaf · PostgreSQL · Docker

A comprehensive invoice management solution built with clean architecture principles and modern Java practices. Collaborative project focused on JPA relationships and system design.

Built with @kristina0x7 and @fmazmz.

Core Java fundamentals in focus — Object-Oriented Design, efficient data handling, and Test Driven Development with JUnit.

Stack: Java · JUnit · JDBC · Testcontainers · Docker