Ayush Shah is a research scientist at Meta with eight years of experience building interpretable, high-performance visual parsers for complex graphical notations such as mathematical and chemical formulas. He completed a PhD at Rochester Institute of Technology working on graph-attention techniques that preserve natural, human-readable graph representations while improving speed and accuracy over typical encoder–decoder approaches. His ChemScraper project demonstrates a pragmatic bent—parsing molecule diagrams from raster PDFs into visual and chemical graphs without OCR, GPUs, or vectorization, and enabling creation of fine-grained annotated datasets. Earlier internships and roles at Amazon and Fusemachines show applied experience scaling semi-supervised speech and NLP/CV systems and shipping production-ready ML pipelines. Based in Menlo Park, he combines deep research rigor with product-focused engineering and a knack for making complex models both fast and interpretable. An often-overlooked strength is his cross-domain fluency—from speech and LLMs to graphical structure recognition—enabling creative multimodal solutions.
8 years of coding experience
6 years of employment as a software developer
Bachelor's degree, Computer Engineering, 3.96 CGPA, Bachelor's degree, Computer Engineering, 3.96 CGPA at Kathmandu University (KU)
Doctor of Philosophy - PhD, Computing and Information Sciences, Doctor of Philosophy - PhD, Computing and Information Sciences at Rochester Institute of Technology
SLC, 88.6%, SLC, 88.6% at Graded English Medium School (GEMS)
Plus 2, Science, 86.2%, Plus 2, Science, 86.2% at GEMS Institute of Higher Education (GIHE)
Contributions:15 commits, 14 pushes, 1 branch 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.