Punit Mehta

Experience

1. Applied Scientist - Machine Learning

   Amazon

   Oct 2019 — Present · Bengaluru, IN

   Primarily into the development of ML models for automatic speech recognition (ASR) in Alexa. Part of a stellar team involved in the design, implementation and deployment of ML (ASR) models for Alexa in English locales.\\\\n\\\\nDeep learning for language modeling in everyday speech recognition tasks, bias in machine learning at the Amazon scale!\\\\n\\\\n* Oct \'19 - April \'20 *\\\\n- I worked as a Machine Learning Scientist Intern with Alexa ASR and was primarily involved in the regionalization - deep learning (TensorFlow & AWS), feature modeling, statistical learning, bias/variance of ASR models. Designed an end-to-end deep learning pipeline from capturing relevant inputs, training models and evaluating for various metrics used for the production models.\\\\n- Established several interesting research directions based on the experiments carried throughout the internship primarily to improve ASR regionally - having a single model to learn all the local nuances. Offered a full-time position based on the internship performance.\\\\n \\\\n\\\\n \\\\n \\\\n\\\\n \\\\n

Education

• Georgia Institute of Technology

  Master of Science - MS, Computer Science (Specialization: Machine Learning), 4.0

  2018 — 2020

• Gayatri Vidhyalaya

  High School, Intermediate/+2

  2010 — 2011

• Stony Brook University

  PhD (on leave), Computer Science

  2017 — 2018

• Dhirubhai Ambani Institute of Information and Communication Technology

  Bachelor’s Degree, Information and Communication Technology (major in Computer Science), 9.15/10

  2011 — 2015