Hi, I’m Joep. I’m a PhD candidate in Delft working on machine learning for computational mechanics. I’m focusing on accelerating physics-based simulations with machine-learning based techniques without sacrificing robustness. During my PhD, I’ve had research collaborations with Columbia and Aarhus University.

Next to my work on computational mechanics, I’ve also been drawn to technical AI Safety. I’ve organized programs at the Delft AI Safety Initiative, and worked on several projects, such as AI Safety Camp and the Dutch AI Safety project competition. I am actively looking for a full-time role to contribute to technical research on aligning AI.

Updates

• 2026-07-22 Talk at World Congress on Computational Mechanics (München): Microstructure-Conditioned Surrogates For Multiscale Material Optimization
• 2026-06-04 Talk at Graphs & Data (Delft)
• 2026-05-14 Paper accepted (3rd author): Predicting plastic strain localization in porous solids using GNNs. In collaboration with Aarhus University.
• 2026-02-15 LLM-evolved digital art project
• 2025-10-30 Poster at Engineering Mechanics symposium
• 2025-10-01 Paper out: Mixing data-driven and physics-based constitutive models using uncertainty-driven phase fields
• 2025-07-01 Talk at CM3P (Porto): Accelerating Multi-Scale Simulations using Uncertainty-Driven Phase-Field Mixtures of Constitutive Models
• 2025-05-22 Assisted organization and presented talk at EuroMech (Gothenburg): A microstructure-based surrogate for multiscale simulations using GNNs
• 2025-04-19 AI Safety Camp project: Mechanistic Interpretability Via Learning Differential Equations
• 2025-04-03 Talk at Graphs & Data (Delft)
• 2025-02-18 Talk at AICOMAS (Paris): Uncertainty-Driven Phase-Field Mixtures of Constitutive Models
• 2024-12-15 Dutch AI Safety Retreat. Individual project (2nd place): Ensemble fine-tuning for pluralistic alignment
• 2024-11-12 Poster at Dutch Computational Science Day (Utrecht)
• 2024-11-01 Talk at DelftBlue supercomputer Summit: Leveraging GPU's for GNNs.
• 2024-07-01 Paper published: A microstructure-based graph neural network for accelerating multiscale simulations
• 2024-06-04 Talk at ECCOMAS (Lisbon)
• 2024-03-01 3-month research visit to Columbia University, collaborating with WaiChing Sun
• 2023-10-13 Attended CMCS
• 2023-09-26 Talk at MMLDE (El Paso)
• 2022-10-22 Attended AI Safety NL retreat
• 2022-07-29 Deep Learning Summer School