Vibhu Agrawal
About Vibhu Agrawal
Vibhu Agrawal is a Tech Lead in Deep Learning with a strong academic background in Electronics and Communications Engineering and Mathematics from the Birla Institute of Technology and Science, Pilani. He has extensive experience in developing frameworks for large language models and has held various roles in data science and deep learning across multiple organizations.
Current Role at NanoNets
Vibhu Agrawal currently serves as a Tech Lead in Deep Learning at NanoNets, a position he has held since 2024. In this role, he oversees the development and implementation of deep learning solutions, focusing on enhancing the capabilities of the company's technology. His responsibilities include leading projects that involve large language models and ensuring efficient training and inference processes.
Previous Experience at NanoNets
Before his current role, Vibhu Agrawal worked at NanoNets as a Deep Learning Engineer from 2021 to 2023. During this time, he contributed to various projects, including the development of an adapter-based serving framework for scalable serving of large language models. He also implemented automated fine-tuning and inference processes aimed at improving document processing efficiency.
Education and Academic Background
Vibhu Agrawal studied at the Birla Institute of Technology and Science, Pilani, where he earned a Bachelor of Engineering (Hons.) in Electronics and Communications Engineering from 2014 to 2019. He also completed a Master of Science (Hons.) in Mathematics at the same institution during the same period. His academic background provides a strong foundation for his work in deep learning and data science.
Research Experience
Vibhu Agrawal has gained valuable research experience through various internships. He worked as a Research Intern at the Indian Institute of Remote Sensing (IIRS) under ISRO in 2017 for two months. Additionally, he served as a Research Scholar at the University of Pennsylvania in 2018 for five months, where he further developed his analytical and research skills.
Technical Contributions
Throughout his career, Vibhu Agrawal has made significant technical contributions in the field of deep learning. He led the development of a constrained decoding framework for large language model inference, ensuring accurate JSON output with predefined keys. He also developed systems for spatially restricted text generation and key-value pair extraction from documents, demonstrating his expertise in advanced machine learning techniques.