Prashant Mathur
Senior Applied Scientist, Amazon Catalog
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolPrashant Mathur is a senior applied scientist with 13 years of experience building production-grade machine translation, speech translation, and multimodal generative AI systems at Amazon, currently improving the shopping experience on amazon.com. He holds a PhD in Computer Science and has led science teams across Amazon projects including Translate, AGI knowledge graphs, and Amazon Q, blending deep research with product-driven deployment. His earlier academic and industry work focused on domain adaptation and online learning for SMT, and he has contributed bug fixes and feature enhancements to the widely used Moses statistical MT decoder, including a cache-based translation model. Based in New York, he combines strong research credentials with practical engineering—bridging core scoring and caching improvements in legacy MT with modern generative and multimodal services.
13 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy (PhD) Computer Science, Doctor of Philosophy (PhD) Computer Science at Università di TrentoMS Computer Science, MS Computer Science at International Institute of Information Technology Hyderabad (IIITH)XIIth, XIIth at Modi Public School, KotaXIth, XIth at St. John's School, KotaXth, Xth at St. Anselm's Pink City School, Jaipur
Hindi, English, Italian
Github Skills (7)
caching10
cachemanager10
c-language10
cprogramming-language10
algorithm9
algorithms9
machine-translation8
Programming languages (6)
Github contributions (5)
moses-smt/mosesdecoder
May 2016 - May 2018
Moses, the machine translation system
Role in this project:
Back-end Developer
Back-end DeveloperContributions:26 commits, 3 PRs, 1 push in 2 years
Contributions summary:Prashant's contributions primarily involve modifications and bug fixes to the core components of the Moses machine translation system. These changes encompass updates to the `GlobalLexicalModel` and the `PhraseDictionaryDynamicCacheBased` modules, suggesting expertise in statistical machine translation and phrase-based decoding. Further contributions include merging updates and implementing features within the CBTM framework to handle additional feature scores, and providing improvements to the core scoring mechanism. The user has also added a new cache based translation model.
i18nmachine-translationtranslationlocalizationmoses
amazon-science/iwslt-autodub-task
Jan 2023 - Mar 2024
Contributions:3 reviews, 2 PRs, 25 pushes in 1 year 1 month
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.