Curriculum Vitae
AI engineer with a Master AI from the University of Amsterdam. I build scalable AI pipelines and predictive models, and I like to learn a field by building the thing that proves I understand it.
Rather than a bullet list, here is the story: how a Bachelor in AI, a part-time job at a shipyard, a research-heavy Master, and a full-time engineering role each taught me a different layer of the craft. The skills below are not claims, they are what the projects forced me to learn.
Download PDF CVBSc Artificial Intelligence
Where the foundations were poured. Predicting where ambulances would be needed from 84k emergency calls taught me time-series modelling and the honest limits of LSTMs and Random Forests on messy data. An NLP system for a wine startup showed me a model only matters once it becomes a product. A lattice-Boltzmann fluid simulation kept me close to the computational science under the abstractions, and a deepfakes essay kept ethics in view alongside accuracy.
Damen Dredging Equipment
Alongside my studies, where AI stopped being academic. I engineered predictive models for maritime dredging equipment, driving maintenance optimisation and precise parameter estimation. The core lesson was validation: bridging a clean research algorithm and messy real-world shipyard data in MATLAB, C++ and Python, training CNNs that had to survive contact with industry. It culminated in my Bachelor thesis, done on-site.
MSc Artificial Intelligence
A deliberate deep dive into research. On the deep learning side I extended a rotation-equivariant graph neural network for proteins, detected rib fractures in CT with 3D medical imaging, and shipped a from-scratch CNN to Google Cloud Vertex AI, learning the MLOps that turns a model into a service. A run of NLP research, probing what transformers versus LSTMs actually know, measuring bias, scoring hallucination, sharpened my instinct for rigorous evaluation. Threaded through it all was the brain, ending in a Master thesis on predictive coding as an alternative to backpropagation at the SILS lab.
Alva Minds
Where the research training met production. I architect end-to-end AI pipelines that pull intelligence from unstructured data and own the full lifecycle on Azure: autonomous agents with LangChain, retrieval over vector indexes and RAG, secure MCP servers, and a React front-end. A recurring theme is squeezing more from less, optimising LLM inference so systems scale. The same instinct spills into my own time: a multilingual RAG chatbot, a self-hosted home lab with Kubernetes and Spark, and reproducing SOTA cell-nucleus segmentation with a swarm of small local LLMs on a GPU-less server. Increasingly the work is agent orchestration, driving Claude Code and other harnesses to move faster without losing rigour.