Vikas Raunak
Senior Research Scientist, Cloud + AI, Microsoft

150th Avenue, NE
Redmond, WA
viraunak@microsoft.com
I am a Research Scientist in the Machine Translation Team at Microsoft. I received my MS in Language Technologies at the School of Computer Science, Carnegie Mellon University, advised by professor Florian Metze. Prior to that, I graduated from IIT Indore, India with a B.Tech in Computer Science & Engineering and the President of India Gold Medal.
My research interests are mainly in Sequence Modeling, Evaluation and Representation Learning with a strong focus on large-scale text and speech (natural language) applications. Specifically, I am interested in:
-
Building computational models which could capture rich sequence-level phenomena (e.g., compositionality) at different scales.
-
Constructing compact representations & systems which could allow accessible manipulations at different levels of abstraction.
The overarching mission of my research is to build scalable, efficient and trustworthy natural language processing (NLP) models & applications which could be reliably used in rich real-world settings.
news
Oct 2023 | Papers on Dissecting In-Context Learning + Post-Editing accepted to EMNLP Findings |
May 2023 | MT-GPT paper is out on arxiv, LLM Literalness paper accepted to ACL main conference |
Oct 2022 | Three papers accepted at EMNLP: Finding Memo, SALTED and GEM version 2 |
Jul 2022 | Invited Talk at IST & Unbabel Seminar: Slides, Check out the Related Blog Post |
Jun 2022 | Papers on SALTED, GEM version 2, Big-Bench and NL-Augmenter are on arxiv |
Mar 2021 | Papers on Hallucinations and Searchable Hidden Intermediates accepted at NAACL |
Aug 2020 | Papers Accepted at ACL Repl4NLP Workshop, CIKM, EMNLP Findings |
Mar 2020 | Serving on ACL-SRW, ALVR, EMNLP, AACL 2020 Program Committees |
Oct 2019 | Giving a talk at CMU LTI Student Research Symposium |
Sep 2019 | Papers on Visually-Grounded NMT at INLG, Compositional NMT at a NeurIPS Workshop |