I’m a researcher at Amazon AWS AI, where I currently contribute to Amazon Quick - an agentic AI assistant for work, previously on Amazon Q - an enterprise conversational assistant, and Amazon Translate - custom translation and automatic dubbing. I hold a PhD from University of Trento & FBK Institute, where my thesis focuses on multilingual and custom neural machine translation.
Research Focus
- 🛡️ AI Safety / Responsible AI (RAI)
- 📊 Generative AI Evaluation (Agentic Eval)
- 🔍 Retrieval-Augmented Generation (RAG)
- 🌐 Neural Translation, Multilingual NLP & Automatic Dubbing
News & Updates
- 2025–present 🚀 Research contributions to Amazon Quick — an agentic AI assistant for work.
- 2024 🚀 Research contributions to Amazon Q — an enterprise conversational AI assistant.
- Sep 2023 📣 Program Co-Chair of the 1st International Workshop on Human-in-the-Loop Applied Machine Learning (HITLAML), University of Luxembourg.
- Aug 2023 📄 Paper Jointly Optimizing Translations and Speech Timing to Improve Isochrony in Automatic Dubbing at Interspeech 2023.
- Sep 2022 📄 Paper Isochrony-Aware Neural Machine Translation for Automatic Dubbing published at InterSpeech 2022.
- Jul 2022 📄 Paper Sockeye 3: Fast Neural Machine Translation with PyTorch released.
- May 2022 📄 Paper Isometric MT: Neural MT for Automatic Dubbing published at ICASSP 2022.
Blog & Notes
Research & Publications
Jointly Optimizing Translations and Speech Timing to Improve Isochrony in Automatic Dubbing
arXiv:2302.12979
2023
Paper
Self-Learning for Zero-Shot Neural Machine Translation 🏆 Best Paper
LoResMT @ MT Summit 2021
2021
Paper
Multilingual and Customized Neural Machine Translation Thesis
PhD Thesis, University of Trento & FBK HLT-MT Lab | Advised by Marcello Federico
My doctoral thesis investigates approaches to improve the performance of neural machine translation models for language pairs without large-scale training data. This work explores zero-shot translation, verbosity control, and isochrony-aware models for automatic dubbing.
Improving Zero-Shot Translation of Low-Resource Languages 🏆 Best Paper
IWSLT 2017
View all research & publications →
R&D Projects
-
🎬Automatic Dubbing — A research hub tracking progress on automatic dubbing — isometric and isochrony-aware translation, verbosity control, and speech-timing models.
Community Contributions
- Program Committee / Reviewer 2017–*: ACL, IWSLT, EMNLP, COLM, ICLR, ICML, NeurIPS.
- Mentorship: Mentoring junior researchers and students in Neural MT & NLP.
- Open Source: see R&D Projects section.