Nitin Madnani
Staff AI Research Scientist at ASSOCIATION FOR COMPUTATIONAL LINGUISTICS INC
New Jersey, United States
Summary
👤
Senior🎓
Top SchoolNitin Madnani is a Staff AI Research Scientist with roughly 14–15 years of experience building and deploying NLP and ML systems in industry and research settings. He currently leads AI research at Duolingo after architecting shared ML/NLP infrastructure and automated scoring systems as a Distinguished AI Engineer at ETS. His work spans practical productionization—IaC on AWS, distributed message-based scoring services—and core research in paraphrase modeling, grammatical error correction, and automated feedback for language assessment. An active open-source contributor and "rabid Pythonista," he has improved model explainability tooling through contributions to the widely used LIME library, emphasizing both reliability and usability. He also volunteers as CIO for the Association for Computational Linguistics, blending operational stewardship with ongoing research leadership. Trained with a PhD in computer science from the University of Maryland, he combines deep academic foundations with a track record of turning NLP research into scalable products.
14 years of coding experience
23 years of employment as a software developer
Ph.D, Computer Science, Ph.D, Computer Science at University of MarylandBE, Electrical Engineering, BE, Electrical Engineering at Punjab Engineering College
Hindi, Sindhi
Stackoverflow
1reputation
50reached
0answers
1question
Github Skills (9)
machine-learning10
explainable-artificial-intelligence10
python10
scikit-learn9
scikit9
testing8
documentation7
javascript6
database6
Programming languages (23)
Github contributions (5)
marcotcr/lime
Jun 2017 - Jun 2017
Lime: Explaining the predictions of any machine learning classifier
Role in this project:
ML Engineer
ML EngineerContributions:5 commits, 18 comments, 4 issues in 1 day
Contributions summary:Nitin made several commits related to the Lime (Local Interpretable Model-agnostic Explanations) library, suggesting a focus on model explainability. The contributions include incorporating regression changes, fixing import errors, and ensuring proper propagation of parameters, likely to refine model explanation functionality. They also refactored testing procedures and updated documentation, indicating a focus on both functionality and usability improvements of the library.
classifiermachine-learning
frcchang/zpar
Jun 2015 - Jul 2015
ZPar statistical parser. Universal language support (depending on the availability of training data), with language-specific features for Chinese and English. Currently support word segmentation, POS tagging, dependency and phrase-structure parsing.
Contributions:2 releases, 19 commits, 8 PRs in 15 days
parserparsingtraining-dataword-segmentation
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.