Jungi Kim
Senior Applied Scientist (NLP) at Thomson Reuters
San Diego, California, United States
Summary
🤩
Rockstar🎓
Top SchoolJungi Kim is a Senior Applied Scientist in NLP at Thomson Reuters Labs with a PhD in Computer Science and over a decade of experience bridging academic research and production ML systems. He specializes in multilingual and multimodal natural language analysis, machine translation, and information retrieval, having led projects from attentional encoder-decoder models to an American Sign Language-to-English translation system. His background includes postdoctoral work on semantic IR and substantial open-source contributions to the widely used OpenNMT project, where he improved sampling and training efficiency for neural machine translation. Based in San Diego, he combines deep learning research rigor with practical engineering to deliver AI assistants for legal professionals, and he brings an uncommon mix of linguistic toolkit development and hands-on model optimization to applied NLP challenges.
10 years of coding experience
19 years of employment as a software developer
Doctor of Philosophy - PhD Computer Science and Engineering, Doctor of Philosophy - PhD Computer Science and Engineering at Pohang University of Science and TechnologyBachelor of Science - BS Computer Science, Bachelor of Science - BS Computer Science at Illinois Institute of TechnologyParaclete High School
Github Skills (13)
lua10
torch-hub10
machine-translation10
deeplearning-ai10
torch-net10
deep-learning10
neural-machine-translation10
nlp9
pytorch9
algorithms8
data-structures8
algorithm8
data-structure8
Programming languages (4)
Github contributions (5)
OpenNMT/OpenNMT
Jan 2017 - Nov 2017
Open Source Neural Machine Translation in Torch (deprecated)
Role in this project:
ML Engineer
ML EngineerContributions:7 commits, 6 PRs, 2 pushes in 10 months
Contributions summary:Jungi made several significant contributions to the `opennmt` repository, which focuses on neural machine translation. Their primary contributions involved enhancing the `SampledDataset` class, adding data sampling capabilities during training, and optimizing the sampling process. These changes included introducing perplexity-based sampling and refining the batching strategy to improve training efficiency, with specific modifications to data sampling logic and logging within the dataset. Further work included adjustments related to Sentence-level Log-likelihood Criterion.
nlpmachine-translationtranslationtorchnmt
jungikim/OpenNMT-tf
Nov 2018 - Sep 2023
Neural machine translation and sequence learning using TensorFlow
Contributions:33 pushes, 7 branches in 4 years 10 months
sequencedeep-learningmachine-translationtranslationmachine-learning
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.