Niranjan Suthar
@niranjansuthar
Computer Vision Engineer at AVA Retail
Surat, Gujarat, India
Niranjan Suthar is a Computer Vision Engineer with expertise in deep learning, image processing, and object detection. He has experience configuring large-scale training environments using Horovod and TensorFlow, achieving significant efficiency improvements. His background includes developing solutions for bio-metrics identification, image segmentation, and shape estimation, demonstrating proficiency in Python and MATLAB.
Experience
Computer Vision Engineer
AVA Retail
Modified the existing Object Detection algorithm to learn new classes without adding all previous class synthetic data-set using knowledge distillation/LwF techniques for retailing shops. Configured existing Object Detection model to detect larger classes with lower variance by dividing into clusters at multiple stages. Researched and configured 15+ GPU machines to train an Object Detection model parallelly using mpi all-reduce, all-gather and data-parallelism, resulting in reduced training time with efficiency of 65%.
Summer Intern, Product Developer
NRB Bearings
Designed an automatic electrical platform on paper to increase the output efficiency from 2 minutes to 30 seconds. Modified hopper design to increase space utilization by 40 percent and tested with prototype design.
Education
Indian Institute of Technology Bombay
Bachelor and Master of Technology
Electrical Engineering
Licenses & Certifications
Computer Vision with Python
EDX
Deep Learning with TensorFlow
IBM
Computer Vision and Image Analysis
Microsoft
Introduction to Artificial Intelligence (AI)
Microsoft
Essential Math for Machine Learning Python Edition
Microsoft
Image Processing, Computer Vision, Remote Sensing and Image Processing
HarvardX
CS50’s Introduction to Computer Science
HarvardX
CS50’s Introduction to Artificial Intelligence
HarvardX