👋 Hi! My name is Kelvin, I have been working since 2020, on ways to automatically process questions (generating, decomposing and answering ❓❓❓), from text, from knowledge graph, simple to complex. Since May 2026, I have been a research fellow at the Singapore Institute of Technology for an online trust & safety project, under principal investigators Ian McLoughlin and Tong Rong.

I did my PhD (on question generation, naturally) at the Université de Lorraine in the lovely city of Nancy, France under the supervision of Claire Gardent and Thiago Castro Ferreira. My PhD was carried out under Project QUANTUM which was funded by the French National Research Agency (Agence National de Recherche, ANR), and I was part of the Synalp team (now within the MosAIk group) at the LORIA (Laboratoire lorrain de Recherche en Informatique et ses Applications) laboratory.

Before that, I did my MSc (Natural Language Processing) at the Institut des sciences du Digital, Management & Cognition (IDMC) also in the Université de Lorraine. I also hold an MA (South-east Asian Studies) from the National University of Singapore where I focused on Indonesian political economy. I did my BSc (Economics) with a minor in Political Science from the Singapore Management University.

Recently, I’ve been working on question generation useful for discourse representations and on post-training LLMs with reinforcement learning. I am exploring extensions of my research into multimodal settings, especially with the generation of structured representations useful as abstractions of scenes and video. Some of my recent work can be found here:

     ⚒️ decomposing complex questions Here I proposed the use of panels of smaller-sized LLMs to obtain decomposition candidates as well as to select from them via voting (ranking with LLM-as-judge).

     📜 generating Questions under Discussion (QUD) Here I proposed using reinforcement learning (GRPO) to obtain improved question generation that have to meet multiple constraints requiring reasoning over a piece of discourse. QUD is an emerging linguistic framework for representing discourse structure; see here for a natural language processing-focused survey.

     🔢 captcha recognition This work is exploratory and was rapidly prototyped, and the images are in a controlled setting. Here I used GRPO with a small set of exemplars to help improve vision language model (VLM) recognition of ambiguous characters.