Karel van der Weg
IFF
Karel van der Weg is a dedicated researcher with an affection for biochemical data and machine learning. He began his academic journey with a BSc. in chemistry at Utrecht Univeristy, followed by a MSc. in chemical engineering at the University of Twente. He then pursued a PhD focused on machine learning for enzymes, where he developed enzyme databases and advanced machine learning methods, including 3D-aware graph neural networks, multimodal systems and generative models for protein dynamics.
Currently Karel works at International Flavors & Fragrances (IFF), where he combines expertise in artificial intelligence and biochemistry to identify and generate novel lead candidates. He also develops automation systems that enable a fully data-driven, lab-in-the-loop research workflow.
With extensive experience at the intersection onto AI and life sciences, Karel regularly shares his expertise through presentations, lectures and educational activities for both specialist and broader audiences.
Currently Karel works at International Flavors & Fragrances (IFF), where he combines expertise in artificial intelligence and biochemistry to identify and generate novel lead candidates. He also develops automation systems that enable a fully data-driven, lab-in-the-loop research workflow.
With extensive experience at the intersection onto AI and life sciences, Karel regularly shares his expertise through presentations, lectures and educational activities for both specialist and broader audiences.
Shu Zhao, Karel van der Weg
Navigating through the protein engineering landscape with automated benchmarking
From bioethanol to laundry detergents, the biotech industry is developing enzymes that make everyday processes greener. During enzyme development, we engineer specific mutations to obtain desired properties, nowadays, often with the use of machine learning (ML) methods. However, it has become increasingly difficult to track which models perform best for our property of interest. Together with IFF, the Marks lab and Xebia, we developed a community driven framework that makes benchmarking as simple as a pull request. We present two packages to facilitate that process, proteingym-base and proteingym-benchmark. In proteingym-base we standardize the collection of biochemical data modalities into a single data-science ready archive, while proteingym-benchmark provides the setup for automated model benchmarking.
Woensdag 12.30 - 13.00 • Lezingenzaal 8
Terug naar het overzicht
Datagedreven leiders ontdekken essentiele kennis, contacten en technologie tijdens Data Expo
Of je nu al een vergevorderde data-strategie hebt ontwikkeld of je organisatie juist de eerste stappen wil zetten richting een datagedreven onderneming; Data Expo reikt jou de oplossingen om de volgende stap te zetten.
We believe data drives digital transformation
Blog
De kracht van Retrieval-Augmented Generation (RAG) ontsluiten
Digitale Transformatie voor MKB: 8x Voordelen en Uitdagingen
naar boven