Technical Experience & Education

Research Internship — Amsterdam UMC

Master's thesis project on deep learning for dynamic MRI reconstruction. The core challenge: MRI acquisition is slow because of fundamental physics constraints — the goal is to learn how to reconstruct good-quality images from far fewer measurements, making scans faster without clinical compromise.

• Technical scope: Python, PyTorch, custom dataset pipelines, model architecture design, quantitative image evaluation (SSIM, PSNR, perceptual metrics), and careful analysis of failure modes
• Research discipline: literature review, ablation experiments, reproducible training runs, and written thesis communication to both technical and clinical audiences
• Domain context: working with medical imaging data requires extra rigour — patient data, clinical relevance, and the gap between benchmark performance and real-world behaviour

Trainee — Province of North Holland

Worked on Data & AI projects in a public-sector environment where technical decisions have real consequences and non-technical stakeholders need to understand and trust them. A useful counterweight to pure research: the problem framing matters as much as the method.

• Scoped and contributed to analytical projects with organisational relevance
• Strengthened the skill of translating data work into plain language — and pushing back when a question is unanswerable with available data
• Practical exposure to the responsible AI questions that get skipped in academic settings: bias, auditability, explainability under political scrutiny

Secretary, AI Study Committee — University of Amsterdam

Represented MSc AI students in the programme's study committee. Coordinated meetings between students and faculty, drafted agendas and minutes, and advocated for curriculum improvements based on student feedback. Kept a committee with competing views moving forward.

Physics Teacher — Zaanlands Lyceum

Designed and delivered physics lessons for HAVO and VWO students. Teaching is one of the best tests of understanding: if you can't explain it clearly to someone starting from scratch, you probably haven't fully understood it yourself. That principle still shapes how I write code, documentation, and presentations.

• Translated dense technical content into structured, accessible explanations
• Managed classroom dynamics and adapted to different learning speeds in real time

Self-hosting & Homelab

Built and maintain a personal Ubuntu server running Docker containers, nginx reverse proxy, Cloudflare tunnel and DNS, Minecraft hosting, and a GitHub Actions deploy pipeline. This site runs on it. Everything breaks eventually — which is how you actually learn networking, systemd, and when to just restart the container.

• Containers, volumes, compose files, port forwarding, and Cloudflare SSL/TLS configuration
• Monitoring, uptime tracking, and backup automation
• DNS record management, Cloudflare proxying vs. tunnel trade-offs