As a Senior Software Scientist at Thermo Fisher Scientific, I am currently working in the material science department, and am responsible for the implementation and improvement of scientific algorithms (EELS, EDS, etc.) supporting Velox.

De STOK BOX is een speelse activiteiten box voor kinderen tussen de 2-8 jaar. De box geeft handvatten om te observeren, te leren en te creëren door middel van thema's. Elke box heeft een ander thema en inhoud die aansluit bij de leeftijdsgroep! Gieter box: 2-4 jaar (rood) Citroen box: 4-6 jaar (groen) Bril box: 6-8 jaar (blauw)Working mostly on the technical side of the business, such as building the product page and webshop.

Took up more senior responsibilities, being part of a core computer vision competency team, and working on a big project in collaboration with design and doctors from Catharina hospital for bringing cutting-edge segmentation models to production for aiding medical diagnosis and help save lives. Furthermore, pushing research in the direction of latest developments and proposing new projects using e.g. NeRF-like architectures in medical imaging and promoting cross-departmental efforts. Also helping to bring the latest one-short 2D to 3D conversion algorithms in production for helping patients to choose appropriate respiratory masks, clothing, etc. As with previous work, we make heavy use of years of development of a high resolution parametric model of head and body.

Worked on the Bodypass subsidy project where I contributed to the creation of new al-gorithms, including landmark-driven skinning, the segmentation of organs and fat from CT scans, and other geometric algorithms. Furthermore, I was involved in designing the Bodypass architecture for Philips. Beyond Bodypass, I have implemented the live visualization of Philips Dreamwear masks on customers faces, using advanced face track- ing techniques. On other projects I’ve worked, I have effectively applied the latest AI algorithms such as differential rendering and neural networks, including GANs, CNNs, RNNs, etc. to solve complex problems such as the detection of meals in pictures, the generation of daily human activity, skin reverse engineering from spectral data, and cut- ting edge medical image segmentation. Furthermore, I’ve worked on developing motion tracking algorithms from 2D and 3D video using advanced techniques such as differ- entiable rendering and depth estimation from 2D video. During this time, I’ve been the author of several Invention Disclosures (ID) and Technical Notes (TN). I have also supervised students and interns.

Worked on ING’s main pricing engine, allowing me to learn the pricing of various products and get familiar with the codebase. During this period I had to discuss with trades and validators, and come up with models and implement pricing and sensitivity calculations for total return swaps and long-dated forward bonds. I also parallelised and refactored big parts of the code, improving the code quality and leading to big performance improvements. I also worked on the integration of code coverage tools, designing a versioning system for the regression test data. Languages I worked the most with include C++, Python, and CUDA.

My PhD thesis with title "Order from Disorder" was conducted under the supervision of Prof. Dr. Ir. Marjolein Dijkstra. During my PhD, I built cutting-edge Monte Carlo and event-driven molecular dynamics simulations for various models, including a lattice model of binary solvent mixture with colloids, a binary solvent mixture with salt (i.e. including electrostatics), and hard particle models. All simulations where programmed by myself in C and C++, and Python was used for scripting purposes. MPI and Cuda was used for massive parallelizations that were run on the Dutch supercomputer cluster "Cartesius". During the period of my PhD, I supervised three M.Sc. students in collaboration with my colleague, and was teaching assistant in our University's simulation course for three years. As a perk, I greatly improved my skill in foosball during this time!

My Master thesis, titled "Glassy Dynamics of Convex Polyhedra", consisted of building cutting-edge Monte Carlo simulations of convex polyhedral particles and consequently characterizingthe dynamics of the glass phase exhibited by these particles. The simulations where programmed in C and C++, and Python was used for scripting purposes, while I also got very familiar with OpenGL which was used for visualization of the simulation results.