Devamanyu Hazarika
Research Scientist at Meta
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolDevamanyu Hazarika is a research scientist in the Bay Area specializing in post-training and alignment of large language models, currently working on the Llama Research team at Meta. With a PhD from NUS and a decade of experience across Amazon AGI and Alexa AI, he was a core developer of Amazon's Nova model family and co-led its alignment team. His work spans conversational AI, multimodal dialogue understanding, emotion recognition, and controllable empathetic response generation, reflected in 40+ publications at top conferences and frequent program committee roles. Beyond model development, he contributes open-source implementations for conversational emotion recognition and brings a rare combination of production ML engineering and rigorous academic research to model alignment challenges.
10 years of coding experience
4 years of employment as a software developer
Bachelor of Technology (B.Tech.) Computer Science and Engineering, Bachelor of Technology (B.Tech.) Computer Science and Engineering at National Institute of Technology WarangalDoctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at National University of Singapore
English, Hindi, Assamese
Github Skills (15)
pytorch10
eval10
machine-learning10
lstm10
rnn-model10
tensorflow10
trainings10
emotion-recognition10
evaluation10
python10
natural-language-processing10
modeling10
n10
conversational-ai10
implement10
Programming languages (6)
Github contributions (5)
declare-lab/conv-emotion
Nov 2018 - Feb 2020
This repo contains implementation of different architectures for emotion recognition in conversations.
Role in this project:
ML Engineer
ML EngineerContributions:27 commits, 4 PRs, 35 pushes in 1 year 3 months
Contributions summary:Devamanyu implemented and modified several machine learning models for emotion recognition in conversations, specifically focusing on the ICON and CMN architectures. Their contributions involve defining model structures, building input pipelines, and setting up training and evaluation loops within a TensorFlow environment. The user also updated the codebase to be compatible with Python 3.6 and made minor fixes to existing model components.
emotion-recognitionnatural-language-understandingdialogue-systemssentiment-analysistensorflow
devamanyu/cv
Mar 2019 - Jun 2022
Contributions:10 pushes, 7 branches in 3 years 3 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.