Omkar Prabhune

Position title: M.S. Student

Email: oprabhune@wisc.edu

Biography

I am a first-year master’s student in Electrical and Computer Engineering at UW-Madison. I am broadly interested in Machine Learning, Data Science, Embedded Systems, and Software Development. I completed my Bachelor’s in Electronics and Telecommunication with a minor in Computer Engineering from College of Engineering, Pune. I worked with Citi for 2 years as a Machine Learning and Software Engineer.

Education

  • University of Wisconsin-Madison | August 2021 – Present
    M.S. in Electrical and Computer Engineering
  • College of Engineering, Pune | May 2019
    B.Tech in Electronics and Telecommunication
    Minor in Computer Engineering

Research Interests

  • Machine Learning
  • Data Science
  • Computer Vision
  • Embedded Systems

Awards

  • Design Automation Conference (DAC) Young Fellowship, 2021
  • Best Outgoing Student Award by College of Engineering, Pune from the undergraduate batch of 2019
  • Best Student Project Award 2019 for undergraduate theses by Tata Consultancy Services
  • Smart India Hackathon 2019 National Winner, conferred by the Government of India

Publication

Work Experience

University of Wisconsin-Madison

Teaching Assistant – ECE 315 | Fall 2021

  • Conducted lab sessions on EDA tools for PCB design, discussion sessions and grading.

Citicorp Services India

Technology Analyst: Data Science and Software Development | August 2020 – July 2021

  • Built Data Analytics and Visualization dashboard providing insights into large data on trade settlements. Constructed a streamlined data pipeline. Improved user experience by making it intuitive to navigate and comprehend.
  • Developed Natural Language Processing (NLP) engine for entity and intent recognition. Achieved 22.7% increase in accuracy and 350% increase in user engagement in Smart Search Application.
  • RESTful APIs and website development for Data Lineage Tool using Agile methodology to improve the productivity and collaboration of software development teams in the organization.
  • Developed commentary redaction tool using Natural Language Processing (NLP) to automate the process of proofreading, reducing the commentary’s time-to-market and human errors (Citi’s D10X Hackathon – 2nd position).
  • Pilot Projects: Modeled machine learning-based prediction of fraudulent credit card transactions using clustering, developed deep learning-based signature verification to reduce the processing time for document verification.

Nanyang Technological University (NTU), Singapore

Research Intern at Rapid Rich Object Search (ROSE) Lab – Deep Learning | May 2018 – July 2018

  • Modeled Deep Convolutional Neural Network-based text-paragraph image classification.

Siemens Technology and Services Pvt Ltd.

Research Intern – Intelligence Traffic Systems | July 2017 – February 2018

  • Developed Computer vision-based real-time vehicle detection for Adaptive Traffic Signal Controller resulting in travel-time and emission reduction.

Projects

Machine Learning at the Edge for Precision Agriculture (Ongoing)

Wisconsin Embedded Systems and Computing (WISEST) Lab | Advisor: Prof. Younghyun Kim

  • Developing optimized and compressed Machine Learning and Computer Vision models for cattle recognition, tracking, and behavior analysis that can be deployed in resource-constrained environments.

Leaf Grade Recognition Using Computer Vision

Smart India Hackathon 2019 (Winner) | Advisor: Prof. P. P. Rege

  • Feature extraction of leaf images in HSV color space to extract information on color, ripeness and uniformity. Designed artificial neural network for classification of the extracted features and grading the leaves.

SpeakingEyes: Enabling Paralyzed People to Communicate

College of Engineering, Pune | Advisor: Prof. P. P. Rege

  • Developed computer vision-based app for paralyzed patients to communicate using eye-gaze tracking.

Music Annotation using Artificial Neural Network

College of Engineering, Pune | Advisors: Prof. P. P. Rege, Prof. R. Patole

  • Extracted FFT features from audio signal and reduced their dimensionality. Designed artificial neural network to classify these features into 7 classes and identify the musical instrument in the audio.