Anchit Gupta
Member Of Technical Staff at xAI
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolAnchit Gupta is an NLP and generative AI engineer with a decade of experience, currently driving reasoning research at xAI after six years building large-scale language systems at Meta. He holds an MS in Computer Science from Stanford and a CS bachelor's from IIT Bombay, blending rigorous academic training with product-focused research. At Meta he progressed from search NLP to staff research engineer roles, shipping generative models and production-facing NLP components. An active open-source contributor, he improved Facebook Research’s PyText by adding prediction capabilities, fp16 gradient compression, and MADGRAD optimizer support—practical enhancements that ease training and inference at scale. Based in the Bay Area, he combines deep model expertise with systems-minded engineering to move research prototypes toward robust deployment. He’s equally comfortable diving into optimizer/debugging subtleties as he is shaping model-level reasoning features.
10 years of coding experience
7 years of employment as a software developer
Indian Institute of Technology BombayMaster's degree Computer Science, Master's degree Computer Science at Stanford University
Github Skills (8)
pytorch10
machine-learning10
nlp10
trainings10
python10
natural-language-processing10
modeling10
optimization9
Programming languages (3)
Github contributions (5)
facebookresearch/pytext
Aug 2019 - Mar 2021
A natural language modeling framework based on PyTorch
Role in this project:
ML Engineer
ML EngineerContributions:8 commits, 9 PRs in 1 year 7 months
Contributions summary:Anchit contributed to the PyText framework, a natural language modeling framework built on PyTorch, by implementing new functionalities and optimizing existing features. They added a `predict` function for a new task, enabling model predictions without true labels. They also addressed issues related to RNNG training by skipping tree validation in the metric reporter and fixed a bug in the label weights for the DocModel. Furthermore, they enabled fp16 gradient compression and added MADGRAD optimizer support.
language-modelingpytorch
pairlab/robosuite
Apr 2018 - Apr 2019
Surreal Robotics Suite with VICES IROS19 code: standardized and accessible robot manipulation benchmark with physics simulation and integrated analytical controllers
Contributions:92 commits in 1 year
roboticssimulation
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.