Biography
I am a B.Tech + M.Tech (Dual Degree) student at IIT Kharagpur, currently in my fifth year, pursuing joint degrees in Electronics and Computer Science.
Broadly, my research interests lie in machine learning for image, language and speech processing. Previously, I have interned at KLA India Software, working on Neural Architecture Search for image classification and the Language Technology Group at Universität Hamburg, where I worked on Multimodal Dubbing.
Download my resumé.
Contact me at sahadebjoy10@gmail.com
Interests
- Speech and Language Processing
- Multimodal Machine Learning
- Computer Vision
- Reinforcement Learning
Education
-
B.Tech, M.Tech Dual Degree in Electronics and Computer Science, 2023
IIT Kharagpur
-
CBSE 12th Science, 2018
Kendriya Vidyalaya, R.K.Puram
Experience
Machine Learning Intern
- Leveraged Neural Architecture Search (NAS) for improving Defect Classification
- Experimented with efficient NAS techniques to optimize user-defined cost funtions
- Evaluated architecture transferability to other datasets
Research Intern
- Developed Text-to-Speech systems for dubbing-related applications
- Assessed the evaluation of audio-visual synchrony
- Introduced the first German single-speaker video-speech-text dataset
Research Intern
- Developed Multimodal Deep-Learning models for hateful content detection in memes
- Developed multilingual language classification models to improve hate speech detection in Dravidian languages.
AI and Perception team member
- Developed Computer Vision systems for UAVs
- Deployed ROS-based traversal algorithms on Ardupilot
- Researched on Learning-based robotic control
Featured Publications
(2022).
A Deep Dive Into Neural Synchrony Evaluation for Audio-visual Translation.
In ACM ICMI 2022.
(2022).
Merkel Podcast Corpus: A Multimodal Dataset Compiled from 16 Years of Angela Merkel's Weekly Video Podcasts.
In LREC 2022.
(2021).
Ensembling strategies for Transformer-based Offensive language Detection.
In DravidianLangTech, EACL 2021.
(2019).
Warehouse Management Using Real-Time QR-Code and Text Detection.
In IMAV 2019.
Projects
*
Fast Quadrotor trajectory following for Flipkart GRID-2.0 Competition.
For IMAV-2019, Indoor Aerial Robotics competition held in Madrid, Spain