Christos Baziotis
I am a machine learning researcher and member of the founding team of Samaya AI. We are developing an AI platform that provides research tools and LLM Agents for finance, serving leading institutions like Morgan Stanley and a number of Wall Street hedge funds. My research focuses on retrieval (QA, RAG), planning with reinforcement learning, and evaluation of LLM outputs.
Before joining Samaya, I did my PhD at the ILCC in the University of Edinburgh under the supervision of Barry Haddow and Alexandra Birch. My research focused on multilingual translation, large-scale pretraining, and improving performance in low-resource settings by extracting signal from unlabeled and synthetic data. During my Ph.D., I did internships as Research Scientist at Meta AI (FAIR) in Menlo Park and as Applied Scientist at Amazon AI in Palo Alto.
Before starting my PhD, I was a research assistant working on unsupervised text generation (summarization and translation) at the National Technical University of Athens with Ion Androutsopoulos and Alexandros Potamianos.
News
Mar 13, 2024 | Our paper “When Does Monolingual Data Help Multilingual Translation: The Role of Domain and Model Scale” has been accepted to NAACL 2024! The camera-ready version is available in https://arxiv.org/abs/2305.14124. See you in Mexico City 🇲🇽🍻🌵🌮! |
---|---|
Nov 1, 2023 | I successfully defended my Ph.D. thesis and joined Samaya AI as part of the founding team! |
May 22, 2023 | Happy to share our new paper “When Does Monolingual Data Help Multilingual Translation: The Role of Domain and Model Scale”! You can read it on arxiv: https://arxiv.org/abs/2305.14124. |
Jan 21, 2023 | Our paper “Automatic Evaluation and Analysis of Idioms in Neural Machine Translation” was accepted to EACL 2023 (Main Track)! 🥳 |
Oct 18, 2022 | The preprint from my Amazon internship work “Automatic Evaluation and Analysis of Idioms in Neural Machine Translation” is on Arxiv (preprint)! |