Can Udomcharoenchaikit
Research Fellow at Vidyasirimedhi Institute of Science and Technology (VISTEC)
Bangkok, Thailand
Summary
👤
Senior🎓
Top SchoolCan Udomcharoenchaikit is an NLP researcher and research fellow based in Bangkok with 11 years of experience and over five years of postdoctoral work focused on bias mitigation, representation learning, and robustness for out-of-distribution NLP. He builds robust Thai language resources and core NLP models (segmentation, POS, NER) that withstand OOVs, rare words, and spelling noise, and recently transitioned practical models from Keras to PyTorch for Thai romanization. An active open-source contributor, he maintains Thai localization for Mozilla’s high-profile Common Voice project and improves PyThaiNLP, blending research rigor with hands-on engineering. His background spans mechatronics and big data analytics, giving him a rare combination of systems thinking and linguistic focus that drives reproducible, production-ready NLP tools.
11 years of coding experience
Bachelor of Engineering (B.Eng.), Mechanical, Bachelor of Engineering (B.Eng.), Mechanical at Thammasat UniversityMaster of Science (MSc), Big Data and Text Analytics, 77.0%, Master of Science (MSc), Big Data and Text Analytics, 77.0% at University of EssexUNSW Sydney
Portuguese, Thai, English
Github Skills (14)
pytorch10
machine-learning10
internationalization10
localization10
nlp10
translation10
python10
crowdsourcing9
encoder-decoder9
lstm9
text-processing8
natural-language-processing8
computational-linguistics8
open-data7
Programming languages (8)
Github contributions (5)
PyThaiNLP/pythainlp
Sep 2016 - Dec 2020
Thai natural language processing in Python
Role in this project:
ML Engineer
ML EngineerContributions:1 review, 12 commits, 6 PRs in 4 years 3 months
Contributions summary:Can primarily contributed to the Thai Romanization module by replacing the Keras model with a PyTorch-based encoder-decoder LSTM model. This included modifying the code to load and utilize the PyTorch model, along with related changes to dependencies and file structures. Further contributions included fixing documentation and code style (pep8).
nlpword-segmentationpythonthai-nlp-librarylanguage-processing
common-voice/common-voice
May 2019 - Sep 2019
Common Voice is part of Mozilla's initiative to help teach machines how real people speak.
Role in this project:Localization / Internationalization Specialist
Localization / Internationalization SpecialistContributions:24 commits in 4 months
Contributions summary:Can's commits primarily focus on updating and translating the Thai (th) localization files for the Common Voice project. They are modifying the `web/locales/th/messages.ftl` file, indicating they are responsible for adapting the user interface text for the Thai language. Their work involves translating various strings, phrases, and UI elements to ensure the project is accessible and usable for Thai speakers. The contributions demonstrate a strong understanding of localization processes.
internet-freedommachinesinitiativeopen-datavoice
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.