Skip to content

Wittawat Jitkrittum

About me

Wittawat Jitkrittum

My name is Wittawat Jitkrittum (วิทวัส จิตกฤตธรรม) (). People call me "Witta". I am a research scientist working on LLMs at Google DeepMind, New York. My core work is to design good objective functions for improving LLM training efficiency, and inference efficiency. Some research questions and example projects I worked on are as follows.

I also worked on image generative modeling in the past. Please see the full list of publications. My talks can be found here.

Online presence

 ArXiv  DBLP  Github  Google Scholar  LinkedIn  Orcid ID  ResearchGate  Semantic Scholar  Twitter

Some highlights

My resume in PDF can be found here

Experience
Education
Filed patents
  • US20240311405A1: Dynamic selection from among multiple candidate generative models with differing computational efficiencies (2024)
  • US20240135254A1: Performing classification tasks using post-hoc estimators for expert deferral (2023)
Selected awards
  • ICLR 2025 Outstanding Paper Honorable Mention for "Faster Cascades via Speculative Decoding"
  • Google Tech Impact Award 2024 for work on model routing (awarded to only 10 most impactful projects)
  • Google Research Tech Impact Award Winner 2023 for work on cascade models
  • NeurIPS 2017 Best Paper Award for Jitkrittum et al., 2017
  • Full scholarship for PhD study from the Gatsby Unit (2013-2017)
  • Full scholarship for master's degree from Okazaki Kaheita Scholarship (2010-2012)
  • Runner-up at National Software Contest (NSC), Thailand 2010. Designed a character-level word tokenizer with a binary classifier.
  • Runner-up at National Software Contest (NSC), Thailand 2009. Designed one of the first factoid Thai Q\&A systems based on Thai Wikipedia.
Selected academic services
Co-organizer of regional events

Last update: 2025-09-09