Experts

I'm a data scientist at TNO at the Data Science department and a PhD student at Leiden University. My research focuses on Knowledge Graph Extraction, Enrichment, and Evaluation. I'm passionate about combining natural language processing and knowledge representation techniques.

My work focuses on building and using LLMs in a responsible way, i.e. by taking into account considerations like privacy and bias. Prior to joining TNO I worked as a PhD student and postdoc in linguistics. Because of my scientific background, I am an expert in research and evaluation methods as well as (scientific) writing and presenting.

With a background in semantic mappings and information retrieval, I worked a lot with word2vec and BERT models. Now the majority of my work is using LLMs (and other methods such as (K)RAGs). I mainly focus on Hybrid AI, combining data-driven NLP methods with knowledge-driven methods, and I am keen on good methodology and proper evaluation.

In my research, I focus on reliable, understandable AI systems to assist users in a trustworthy manner on important decisions and analyses. For this, I research hybrid AI approaches by combining Natural Language Processing methods and knowledge driven methods (knowledge graphs, ontologies). Examples of my work are research on conversational recommender systems within the EU-FarmBook project, development of the concept matching tool, or research on knowledge graph and ontology extraction for the TrustLLM project.

I apply and evaluate LLMs for NLP tasks such as classification, question-answering (via Retrieval Augmented Generation (RAG)) and summarization, synthetic data generation and for communicative interfaces. I am working on agentic communicative systems that are able to use tools (vector search, AI models) and external knowledge bases with the aim to enhance human-AI collaboration. My goal is to develop systems that seamlessly integrate with someone's personal- or professional life in such a way that human and machine primarily do tasks that they are good at or enjoy.

My work often focuses on the overlap between NLP and Responsible AI. I like to take into account both the technical perspective as well as a broader organizational or ethical perspective. For example, I have been involved in Impact Assessments of AI systems, have investigated how LLMs can be evaluated technically, and have performed experiments with users of LLM systems.

I specialize in building Large Language Models and adapting them for applications in various projects. My research focuses on demonstrating the feasibility of applying LLMs and agent technology to new applications.

I like working on the practical aspects of language technology, such as examining its reliability, mitigating potential harmful effects, and finding ways to leverage it for social good.

NLP enthusiast, Linguistic and Ontologies aficionado. My work mostly revolves around LLM-integrated frameworks such as RAG chains or Agents based applications. I am interested in unsupervised representation learning, foundational models and their combination with structured human knowledge.

My work focuses on integrating my expertise in AI systems (including LLMs, Computer Vision, and Traditional Machine Learning) with software development to create practical solutions that address critical societal issues such as sustainability and misinformation. I like developing novel approaches to leverage technology in tackling these challenges more effectively.

My work focusses on combining both data-driven and knowledge driven AI and NLP methods. As such I combine Knowledge Graphs and (Large) Language Models to get the strengths of both types of systems for the specific use-case. For example, using LLMs for ontology matching and using a Knowledge Graph with Reasoning and an LLM for a chatbot. My goal is finding the best combination of AI methods for your application.

I work as a consultant and business developer and specializes in language and speech technologies. I apply my knowledge of NLP and ASR in various projects, and am active in a wide range of consortia and networks to promote sovereign and high-performing language models from Europe. I am the product owner of the Dutch LLM initiative GPT-NL and co-lead of the BDVA's Task Force Data & AI Technologies.

After experiencing projects in a range of TRL, I enjoy collaborating with colleagues and clients to design and build NLP related systems for use in a real-world settings, thus not stopping when a model is implemented, but thinking further in how it can be applied in different contexts (For example, would our creation still work when thousands of users use the system at the same time?). Moreover, I enjoy furthering my understanding in LLMs, for example through the creation of evaluation and benchmarking metrics to compare different models. Thus I am interested in both academic as commercial projects and I'm always open for a chat.