I’m a Senior Applied Scientist at Amazon in Barcelona, where I have worked since 2019. Most recently I led the science for bringing customer review insights into the AI Overviews in Amazon search. Before that I worked on the science behind the AI-generated review highlights shown on Amazon product pages. I also develop LLM-as-a-judge methods for evaluating LLMs, used both internally and in research (see my EMNLP 2026 papers below), and mentor applied science interns.
Before Amazon I worked in NLP research. I was a researcher at ISTI-CNR in Pisa, where I also did my PhD with the University of Venice, and then a postdoc with Ivan Titov at the University of Amsterdam, including a visiting period at the University of Edinburgh. More on my About page and on LinkedIn.
Selected work
- AI-generated review highlights
The "Customers say" summaries of customer reviews on Amazon product pages, in many languages and Amazon marketplaces worldwide. - AI Overviews in Amazon search
Bringing customer review insights into the AI Overviews in Amazon search. - Post-hoc Alignment of LLM-Judges to Human Judgment Distribution
Sebastian Steindl, Nikos Voskarides, Alberto Gasparin, Diego Marcheggiani. Proceedings of EMNLP, 2026. - Does Task Decomposition Improve Automatic NLG Evaluation?
Sebastian Steindl, Nikos Voskarides, Alberto Gasparin, Diego Marcheggiani. Proceedings of EMNLP, 2026.