Shrey Desai is a software engineer in Seattle with 11 years of experience building AI-driven products, currently focusing on personalized AI agents for SMBs at Meta. He has a strong ML systems background from contributions to Facebook Research’s PyText—fixing model bugs, improving language-model metrics, and enabling Caffe2 export—alongside internships at Facebook, Uber, and Microsoft. Comfortable moving between research and production, he implements practical model improvements (dropout logic, causal convolutions, GELU activations, pooling) that make NLP models more robust and deployable. A UT Austin computer science alum, he pairs academic rigor with hands-on engineering and a knack for shipping features that bridge prototype research and scalable production.
11 years of coding experience
1 year of employment as a software developer
Bachelor's degree, Computer Science, Bachelor's degree, Computer Science at The University of Texas at Austin
A natural language modeling framework based on PyTorch
Role in this project:
ML Engineer
Contributions:48 commits, 62 PRs, 3 issues in 2 years 4 months
Contributions summary:Shrey primarily contributed to the development and improvement of the PyText framework, a natural language modeling framework based on PyTorch. Their work included fixing weight tying bugs in the LMLSTM model, implementing and improving language model metric reporting, and enabling Caffe2 exporting for LMLSTM models. They also added features like dropout conditions, created causal convolutions, implemented gelu activations, and added pooling mechanisms for CNNs.
A natural language modeling framework based on PyTorch
Contributions:110 pushes, 59 branches in 1 year 10 months
pytorchnlpbertmachine-learningnatural-language
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.