đź‘ľHi, I’m Yiannis (Γιάννης)*. I’m a Research Scientist in the Omnilingual team at FAIR, part of Meta Superintelligence Labs, in Paris. I work on evaluating and improving the multilingual capabilities of AI models. The goal is simple to state and hard to reach: AI that works for everyone, in every language, including the thousands of languages that today’s systems still can’t handle.
I joined Meta full-time in October 2026, where I also did two internships in 2024–2025. My work during the internships led to Omnilingual SONAR and Omnilingual MT: embeddings and translation for thousands of languages. Omnilingual MT received an Outstanding Paper Award at EMNLP 2026.
Before that, I did my PhD at UPC Barcelona on speech translation, and along the way I did research internships at Apple, Dolby and Zeta Alpha, after starting out in quantitative finance in Amsterdam.
*Γιάννης (Yiannis) is the everyday form of Ιωάννης (Ioannis), the name I publish under.
Outstanding Paper Award at EMNLP 2026. We present Omnilingual MT, the first MT system benchmarked on more than 1,600 languages, where 1B–8B parameter specialized models match or …
We introduce OmniSONAR, a family of omnilingual cross-lingual and cross-modal sentence embedding models spanning 4,000+ language varieties and natively supporting text, speech, …
We propose a character-based translation model to improve adaptability to new languages and modalities, particularly for low-resource scenarios. Our method achieves …
We introduce ZeroSwot, a zero-shot speech translation method that aligns a speech encoder with a multilingual MT model using only ASR data, achieving state-of-the-art results …