Scalable backends and interactive systems.
CS graduate (Cum Laude) and Software Engineer at Carecogni. I work across Python backends, Dagster ETL, and React frontends.
Experience
Professional Experience
Software Engineer
Carecogni · Utrecht, Netherlands
Aug 2026 – Present
- Full-time software engineer on Carecogni's healthcare data platform (hybrid, Utrecht).
- Backend: Python, Dagster ETL pipelines, SQLAlchemy and Alembic on PostgreSQL, Pydantic, Poetry.
- Frontend: React.
Software Engineer Intern
Fujifilm Electronic Materials · Zwijndrecht, Flemish Region, Belgium
Jul 2026 – Aug 2026
- Built a local, on-premises RAG assistant for internal documentation, fully offline with no data leaving the building. Used Ollama and Qwen3 for inference and Qdrant as the vector store for retrieval.
- Designed and automated a CI/CD pipeline from git to IIS, replacing manual deployments.
Software Engineer Intern
Utonomy · Utrecht, Netherlands
Apr 2026 – Jun 2026
- Designed and built two production-deployed MCP servers for Utonomy's Solid-based data platform.
- hive-mind: remote MCP server giving AI coding agents (Cursor, Claude Code) shared architectural context of the Utonomy monorepo; service documentation tooling, lookup and search tools, bearer auth middleware; deployed on Scaleway serverless containers.
- u-mind: privacy-first personal context MCP server exposing user pod data (Solid/RDF) to AI clients via OAuth2 + PKCE; consent-aware read/write of personal data models (profile, shopping preferences) with automatic audit logging through idp-server's existing ACL and DataAccessLog infrastructure.
- Tech stack: TypeScript, Node.js, Fastify, MCP SDK (Streamable HTTP), Solid/RDF, Zod, pnpm/Turborepo, Docker, Scaleway, GitHub Actions CI/CD, Linear.
Software Engineer & IT Assistant
Fujifilm Electronic Materials · Belgium
Feb 2025 – March 2026
- Built AI Shark for the Fujifilm Global AI Challenge: internal phishing simulation platform targeting 10 departments; used Fujifilm's internal LLM to generate personalised phishing emails with configurable difficulty, open/click tracking, and admin analytics dashboard. Stack: React, TypeScript, ASP.NET Core 10, Entity Framework Core, SQL Server.
- Developed Asset Management System (AMS) tracking 500+ IT assets with serial numbers, borrowing/returning workflows, and AD-backed role-based access control. Stack: .NET 9, Blazor Server, SQL Server, Dapper.
- Maintained a mission-critical C# Blazor app for the Infrastructure Team, resolving 15 bugs and shipping new features; created PowerShell automation scripts for Cisco switch upgrades across 30 devices.
Education
Bachelor of Applied Science, Applied Computer Science
Karel de Grote University College · Belgium
Class of '26
- Graduated Cum Laude, June 2026.
- Coursework in software development, databases, IT infrastructure, data science, machine learning, and cloud technologies.
- Practical experience across full-stack engineering, business collaboration, and problem-solving through project-based modules and international study.
- Activities: Tutoring — Programming Fundamentals.
Skills in Context
Grouped by how they are used in real projects, not just a logo wall.
Backend
REST and event-driven APIs with Python, Java, Spring Boot, and .NET. ETL with Dagster and SQLAlchemy.
Frontend
Responsive, accessible interfaces with React, TypeScript, and modern CSS frameworks.
DevOps
Cloud infrastructure, CI/CD, and containerized deployments on Scaleway, Azure, and Vercel.
Data & Machine Learning
Datasets, model training, and ML pipelines with Python, scikit-learn, and transformers.
AI & LLM
Production MCP servers, on-prem RAG assistants, multi-agent systems, and privacy-first LLM tooling.
Asset Management System
Asset lifecycle platform for Fujifilm Europe's IT inventory and infrastructure visibility.
Problem
Fujifilm Europe needed a centralized system to catalog and manage IT hardware assets across multiple locations, with unique serial tracking, automated discovery, and full transaction history.
Constraints
- Must integrate with existing Active Directory infrastructure
- Deploy within secure Fujifilm internal network
- Support role-based access for admins, engineers, and end users
- Handle large-scale asset inventory with performance requirements
- Maintain enterprise security standards and compliance
Stack & Architecture
Built with .NET 9.0 and Blazor Server for real-time updates and server-side rendering. SQL Server handles transactional data and asset relationships. Active Directory integration provides authentication and authorization. The system implements a layered architecture with clear separation between data access, business logic, and presentation layers.
Results
- Centralized visibility into IT infrastructure assets across Fujifilm Europe
- Automated asset discovery reduces manual data entry overhead
- Role-based access control ensures proper security boundaries
- Transaction history provides audit trail for compliance
- Shipped into Fujifilm Europe IT operations with role-based access and an audit trail
More Projects
Backend APIs, frontends, ML, and cloud work beyond the featured case study.
Carecogni Platform
Software Engineer
Healthcare data platform: Dagster ETL, SQLAlchemy on PostgreSQL, Pydantic models, and a React frontend. Packaged with Poetry.
On-premises AI Assistant
Software Engineer Intern
Local RAG assistant for Fujifilm internal docs. Fully offline with Ollama, Qwen3, and Qdrant so nothing leaves the building.
Utonomy MCP Servers
Software Engineer Intern
Two production MCP servers for Utonomy's Solid-based data platform: shared architectural context for AI agents, and privacy-first access to personal pod data.
BanditGames
Full-stack Developer
Microservices gaming platform with containerized services for gaming and AI.
Adullam - Bible Journal App
Mobile Developer
React Native app for Bible journaling with AI insights and multi-language support.
Multi-Agent AI Systems
AI Engineer
Multi-agent systems and reinforcement learning project with a dedicated frontend.
Enterprise Full-stack System
Full-stack Developer
Spring Boot backend and React frontend with automated CI/CD pipelines.
EvaleBike
Full-stack Developer
Azure-native evaluation platform for electric bike test benches.
Deep Learning & RL Research
ML Engineer
Deep learning and reinforcement learning implementations.