About Me
I am an avid researcher in field of Artificial Intelligence, Deep Learning and Computer Vision. My research interests include Object Detection, facial analysis (Detection, Identification, Landmarking etc.), 2D/3D Instance Segmentation, Deep Learning model compression, Self-Supervised Learning and Scene Understanding. I also specialize in compute resource constrained real time inference of Deep Learning based Computer Vision models for various mobile devices like Nvidia AI embedded systems, Apple iphone, drones etc.
I currently work full time as a Lead Research Engineer at an exciting Hollywood AI startup called Flawless AI. I spend my time researching and building core Computer Vision AI systems for the movie/TV series domain at my startup as well as working on personal research projects.
Interests
- Artificial Intelligence
- Computer Vision
- Facial Analysis
- Pattern Recognition
Education
MSc in Electrical and Computer Engineering
Carnegie Mellon University, USA
BSc in Electronics Engineering
Veermata Jijabai Technological Institute, India
Recent News
Sept, 2025:
Awarded U.S. Patent for AI research as first inventor! Link
March, 2024:
Filed 1 US patent
Aug, 2022:
Joined Flawless AI as Research Engineer!
Aug, 2020:
1 paper accepted at ECCV 2020.
Feb, 2020:
1 paper published in WIREs Journal.
Jan, 2020:
1 paper accepted at ICASSP 2020.
Jan, 2020:
Joined Honeywell as Advanced Data Scientist!
Feb, 2019:
Tech article on
MedAL paper published.
Dec, 2018:
Received Best Overall Paper Award at ICMLA 2018.
Experience
Research Engineer
Flawless AI
- Conducted lead research on AI based facial analysis tech (Face detection, Face identification, Face motion tracking etc.).
- Developed core AI systems for character facial analysis for movie/TV show domain.
- Conducted research on data efficient techniques for Neural Face Rendering algorithms.
Machine Learning Research Engineer
Honeywell Robotics
- Researched on Deep Learning architectures for various Computer Vision tasks.
- Facilitated AI application knowledge across various Honeywell Robotics teams.
- Successfully built complex Robotics AI based Perception system software.
Lead Deep Learning Researcher
Cylab, Carnegie Mellon University
- Worked in Professor Marios Savvides’s lab at CMU Cylab
- Led a team of Deep learning researchers and iOS app developers to create a proprietary iOS application for driver drowsiness detection.
- Led real time object detection for edge application based projects funded by US DTRA.
- Researched a novel framework for Knowledge Distillation based Deep Learning based model compression.
Summer Research Intern
Cylab, Carnegie Mellon University
- Interned in Professor Marios Savvides’s lab at CMU Cylab
- Researched on computationally efficient yet accurate Computer Vision architectures for performing various facial biometrics tasks which include face detection, face landmarking and face pose estimation.
Graduate Research Assistant
Carnegie Mellon University
- Worked as research assistant to Professor Asim Smailagic in ECE dept, CMU
- Researched onnovel active machine learning techniques being applied for medical imaging analysis.
- Invented a new active machine learning technique to query the most important unlabeled images that the trained model is uncertain about. Technique was developed in context of training CNN models to detect diabetic retinopathy in eye color fundus images.
Skills
Artificial Intelligence
Computer Vision
Robotics
Image Processing
Research
Python
Pytorch
Tensorflow
Project Management
Docker
Recent Publications
VideoClusterNet: Self-Supervised and Adaptive Face Clustering for Videos
In 18th European Conference on Computer Vision (ECCV) 2024
Online Ensemble Model Compression using Knowledge Distillation
In 16th European Conference on Computer Vision (ECCV) 2020
Attentive CutMix: An Enhanced Data Augmentation Approach for Deep Learning Based Image Classification
In 45th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2020
O‐MedAL: Online active deep learning for medical image analysis
In WIREs Data Mining and Knowledge Distillation
Projects
*
Human Driver Drowsiness Detection
Artifical Intelligence based iOS application for detecting driver drowsiness
Real Time Object Detection on Jetson Xavier
Research on compressing state-of-the-art algorithms for edge AI applications
Forecasting Path from Expert Demonstrations
Research on path forecasting for self driving vehicle using CARLA simulator
Deep Audio-Visual Model for Speech Recognition
Research on infusing attention mechanism for combined Deep Audio-Visual Models
Live site visitor locations