Sameer Jain
Machine Learning Engineer II at Pinterest
Mountain View, California, United States
Summary
👤
Senior🎓
Top SchoolSameer Jain is a Machine Learning Engineer II based in Mountain View with nine years of engineering experience and an MS from Carnegie Mellon’s Language Technologies Institute focused on ML and NLP. He bridges research and product, having shipped on-device NLU for Samsung’s Bixby, built evaluation frameworks for generated text published at ACL Findings, and tuned LLMs for meeting summarization at Supernormal. His research with WWF produced state-of-the-art multilingual classification for low-resource languages using clever prompt pipelines that use <6% of typical training data. He also contributes to open-source speech tooling, adapting ESPnet for Burmese ASR data preparation and tokenization. Comfortable across research, production, and prompt engineering, he brings practical efficiency to NLP solutions that need to work in constrained or multilingual settings.
9 years of coding experience
4 years of employment as a software developer
BITS Pilani, Birla Institute of Technology and ScienceMaster of Science - MS Intelligent Information Systems (Language Technologies Institute), Master of Science - MS Intelligent Information Systems (Language Technologies Institute) at Carnegie Mellon University School of Computer ScienceHigh School Central Board of Secondary Education , High School Central Board of Secondary Education at Amity International School, Noida
Github Skills (7)
data-preprocessing10
speech-recognition10
dataprep10
python10
pytorch8
yaml8
machine-translation5
Programming languages (2)
Github contributions (5)
espnet/espnet
Mar 2022 - Apr 2022
End-to-End Speech Processing Toolkit
Role in this project:
ML Engineer
ML EngineerContributions:2 reviews, 15 commits, 1 PR in 1 month
Contributions summary:Sameer primarily contributed to the data preparation and configuration aspects of the Burmese language ASR model within the ESPnet toolkit. They added data preparation scripts and modified existing ones to include Burmese datasets, and they adjusted the configuration files to reflect these changes. The user also updated dataset URLs and parameters such as the nbpe value for tokenization, and added flags for ssl. These modifications indicate their involvement in adapting the toolkit for a new language and optimizing the model's training process.
speech-recognitionspeech-separationchainerspoken-language-understandingspeech-processing
JainSameer06/espnet
Mar 2022 - Apr 2022
End-to-End Speech Processing Toolkit
Contributions:10 pushes, 3 branches in 1 month
end-to-endspeech-to-textspeech-recognitionspeech-synthesisspeech
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.