Marcio Fonseca (PhD 2020–2024) researched summarisation techniques for long documents, exploring how large language models can generate context-tailored summaries. He is now a Machine Learning Engineer at the Brazilian Chamber of Deputies, designing AI solutions for the legislative domain. PhD start and end years 2020 - 2024 What was your PhD research about? My research focused on summarization techniques for long documents, such as financial reports and scientific papers. Among other things, I explored how emerging large language models could be adapted to generate summaries tailored to different contexts. What motivated you to undertake doctoral study? I have always been fascinated by how intelligence works in humans and machines. After completing my MSc in Cognitive Science at the School of Informatics, I became deeply inspired by the research environment and the wide range of topics available for exploration. Pursuing a PhD felt like a natural next step for me. What was a highlight (one or several) of your time as a doctoral researcher? Having the opportunity to present my work in places such as Abu Dhabi and Thailand was incredibly rewarding. These experiences allowed me to engage with diverse cultures and connect with passionate researchers from around the world. What challenges did you face, and how did you overcome them? Aside from the pandemic, one of the main challenges during the PhD was managing the open-ended nature of research. It allows you to explore exciting and uncharted directions, but not all of them lead to publishable results. Balancing research risks and managing time effectively are crucial skills to develop. What are you doing now career-wise, and how did your PhD prepare you for it? I currently work as a Machine Learning Engineer at the Brazilian Chamber of Deputies, where I design ML solutions for the legislative domain, including agentic systems capable of answering queries about legal documents. My work often involves designing human evaluation experiments to assess the usefulness of these solutions—an area in which I gained substantial experience during my PhD. What’s one key skill or mindset you developed during your PhD that you still rely on today? Being curious and open to different perspectives when solving problems is an essential skill. I’ve always admired how the most experienced researchers tend to be the most open to learning new approaches to challenges they already know well. This is a mindset I continue to apply in both my professional and personal life. On the technical side, my PhD gave me hands-on experience in designing and running experiments using the university’s computing infrastructure. Managing computational resources efficiently, automating experiments, and analyzing results became part of my daily routine. This technical experience has been invaluable in my current role, where it directly informed the procurement and configuration of new equipment to support machine learning workloads within my organization. What advice would you give to someone considering a PhD in Informatics? An academic researcher is, to a large extent, a communicator. Even if your work is excellent, you still need to make an effort to present it attractively among the hundreds of other publications at poster sessions and conferences. It’s also valuable to collaborate and publish with colleagues from the University of Edinburgh and other institutions. Is there anything else you’d like to share with prospective students or the wider community? There are plenty of extracurricular activities and sports clubs at the university that help maintain your well-being during the program. I would highly recommend the sailing club! Related links Marcio Fonseca on LinkedIn This article was published on 2025-12-11