Mail   |   LinkedIn   |   Google Scholar   |   GitHub   |   CV

Last updated: September 2025

AI Master's student at the University of Amsterdam with a focus on causality and machine learning. Currently researching causal bandits, developing frameworks that can learn optimal interventions when both the causal structure is unknown and some variables cannot be directly manipulated, under supervision of Assistant Professor Sara Magliacane and PhD candidate Roel Hulsman.

With a background in Computer Science, I bring a hands-on approach to machine learning research, bridging theoretical advances with practical implementations. My industry experience as a software engineer at Twiki Solutions includes leading data-driven services and developing AI solutions for logistics applications.

My research interests lie at the intersection of causality, machine learning, and real-world applications. I'm particularly drawn to problems that require both theoretical rigor and practical implementation—developing methods that advance our understanding while solving meaningful challenges.

I'm motivated by intellectually stimulating research questions, collaborative academic environments, and the opportunity to contribute to the advancement of machine learning theory with real-world impact.

Reproduced and extended the original paper by Basu et al. (2023), verifying results across missing experiments and expanding the discussion on the weight patterns learned by λ‑equitune. Proposed and implemented equiattention, an extension utilizing an Attention layer to potentially improve performance over λ‑equitune.

Reproduced and extended the ICLR’23 paper by Chhabra et al., refining the codebase, introducing additional metrics and datasets, and exploring novel attack strategies. Conducted comprehensive evaluations, including targeted attacks and an ablation study, to validate original claims and broaden the research landscape.

Investigated the efficiency and reconstruction quality of different supervised and self-supervised methods for Magnetic Resonance Imaging (MRI) reconstruction. Extensively compared a pre-trained Cascades of Independently Recurrent Inference Machines (CIRIM) to various methods including Noise2Recon, SSDU, U-Net, ResNet, and traditional Compressed Sensing approaches.

Developed an Alzheimer’s disease detection framework using deep learning, transfer learning, and explainable AI on 2180 MRI images from the ADNI dataset. Implemented a comprehensive preprocessing pipeline and improved model explainability with Grad‑CAM.

Source: adapted from Thijmen Nijdam's fork of Alexander Timans' fork of Dharmesh Tailor's fork of Leonid Keselman's fork of John Barron's website.