Kalvin Chang - Graduate Research Assistant

Kalvin Chang

Graduate Research Assistant

Berkeley, California, United States

Join Prog.AI to see contacts

Summary

👤

Senior

🎓

Top School

Kalvin Chang is a CS PhD student and NLP/speech researcher at Berkeley with nine years of experience building speech foundation models, ASR systems, and production ML tools. He earned top-ranked Language Technologies MS and a CS BS from CMU, and has first-authored work across ACL, COLING, and Interspeech on phonology, historical linguistics, and bias in self-supervised speech representations. His projects range from training distributed Zipformer ASR models for Chinese dialects at Tencent to uncovering AAVE biases and adding AAVE datasets to ESPnet while at CMU, and he contributes to widely used open-source tools like epitran and ESPnet. Kalvin focuses on improving sample efficiency and test-time adaptation for non-standard varieties, blending linguistics-informed modeling with scalable engineering. He’s equally comfortable writing backend pipelines and dataset automation as he is designing phonetic representation learning, and he secured an NSF grant to push multilingual phone-recognition and in-context speech learning.

9 years of coding experience

7 years of employment as a software developer

Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at University of California, Berkeley

Master of Science - MS, Language Technologies, 4.28/4.33 (Rank: 1/18), Master of Science - MS, Language Technologies, 4.28/4.33 (Rank: 1/18) at Carnegie Mellon University

4.84/4.0, 4.84/4.0 at Glen A. Wilson High School

Github Skills (16)

data-preprocessing10

kaldi10

bash10

text-normalization10

data-pipelines10

data-pipeline10

python10

scripting10

natural-language-processing9

speech-processing9

speech-recognition9

nlp9

setuptools7

machine-learning7

regex6

Programming languages (7)

TypeScriptC++JavaScriptHTMLJupyter NotebookRubyPython

Github contributions (5)

dmort27/epitran

Mar 2022 - Oct 2022

A tool for transcribing orthographic text as IPA (International Phonetic Alphabet)

Role in this project:

Backend Developer

Contributions:52 commits, 30 PRs, 28 pushes in 6 months

Contributions summary:Kalvin primarily worked on enhancing the functionality of the epitran tool. They integrated support for different languages and dialects, including tones for Chinese dialects. The user also addressed compatibility issues by reverting code changes to support Python 3.7 and adjusting the project's Python version requirements. This work involved modifying core files and the project's setup configuration.

phonetic-alphabetipaphoneticinternational-phonetic-alphabetphonetics

espnet/espnet

Apr 2022 - Mar 2025

End-to-End Speech Processing Toolkit

Role in this project:

Data Engineer & Automation Engineer

Contributions:7 reviews, 7 PRs, 18 comments in 2 years 11 months

Contributions summary:Kalvin primarily contributed to the data preparation pipeline for the CORAAL dataset within the ESPnet toolkit. Their work involved creating and modifying scripts for downloading, validating, and preparing the data, including text normalization and generating Kaldi-style segments. They also automated processes by integrating snippet generation and dataset splitting into the data preparation workflow. Furthermore, the user implemented scripts for dataset verification and file existence checks.

speech-recognitionspeech-separationchainerspoken-language-understandingspeech-processing

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.

Request Free Trial