Harsh Trivedi
Research Scientist at Allen Institute for Artificial Intelligence
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolHarsh Trivedi is a research scientist specializing in Natural Language Processing and Machine Learning with 12 years of industry and academic experience, currently at the Allen Institute for AI after a long research stint at Stony Brook University. His work focuses on Natural Language Inference and effective information aggregation for tasks like textual entailment and question answering, blending rigorous research with practical model training techniques. He has contributed important tooling to the AllenNLP ecosystem—adding selective parameter freezing, vocabulary extension, and pretrained embedding support—that improve fine-tuning and deployment workflows for NLP models. Based in Seattle, he also builds AI agents and interactive environments (e.g., AppWorld), showing a knack for turning research ideas into user-facing experiments. Colleagues know him for combining deep theoretical insight with engineering pragmatism, particularly around optimizing model initialization and vocabulary management.
12 years of coding experience
9 years of employment as a software developer
Bachelor's degree Information and Communication Technology, Bachelor's degree Information and Communication Technology at Dhirubhai Ambani UniversityMaster's degree Computer Science, Master's degree Computer Science at Stony Brook University
English, Hindi, Gujarati
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (8)
regular-expression10
pytorch10
nlp10
deep-learning10
python10
natural-language-processing10
fine-tuning10
data-science9
Programming languages (9)
Github contributions (5)
allenai/allennlp
Jun 2018 - Jun 2019
An open-source NLP research library, built on PyTorch.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:29 commits, 27 PRs, 308 comments in 1 year
Contributions summary:Harsh contributed significantly to the AllenNLP research library, focusing on enhancing the framework's functionality, particularly around parameter initialization and vocabulary extension. They implemented features allowing for selective parameter freezing during fine-tuning, utilizing regular expressions for parameter selection. The user also added support for extending vocabularies, incorporating pretrained embeddings, and improved the API for these features. These contributions demonstrate a deep understanding of NLP model training and optimization.
nlppytorchtransformerspythonbert
HarshTrivedi/HinglishSentiment
Nov 2014 - Feb 2015
Contributions:36 commits in 2 months
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.