Onkar Kulkarni

About me

Languages: Python, C, C++, R, SQL, Ruby, HTML, Java, Shell Script, PHP
Courses: Machine Learning, Data Mining, Design and Analysis of Algorithms, Software Engineering, Data Structures
Tools and Software: Matlab, Docker, Git, Vim, GDB, Anaconda, Android Studio, MS Visual Studio, MS Office Suit
Frameworks: TensorFlow, Tableau, Django, Flask, Tkinter, QT, WordPress, Octave,
Data Analytics Tools: Sci-Kit Learn, NumPy, Pandas, OpenCV, Theano, YOLO, Kera’s API, Inception V3
OS: Chrome OS, Windows, Linux, Raspbian, ROS, Unix, Kali Linux, Arch Linux
Other: Ruby on Rails, MongoDB, Arduino, BeagleBone, Raspberry Pi, Lex, YACC, Classification and Regression Algorithms

Amdocs Development Centre India LLP December 2018 – July 2019
Technology Engineer
• Reduced ticket processing time from 18 minutes to 3 minutes by integrating server terminal in chatroom
• Improved uptime of servers by ~1% by automating restarts of hung JVM’s measured in active hours per day
• Upgraded alerting system from shell to python 3.0 of payment critical applications for AT&T
VST Infotech August 2017 – December 2018
Software Developer Intern
• Delivered an Android application to provide a tailored real time financial advice to entrepreneurs
• Designed Random Forest Algorithm to predict trends over inadequate financial history of 3 months
• Addressed potential tree overfitting issues in Random forest by correlating features with number of trees
• Demonstrated positive growth of 4% on net profit after recommended advice in effect over 3 months

Robocon 2018 August 2017 – March 2018
• Formulated novel dynamic programming algorithm to navigate 5 possible moves of robot to decrease task time by 1.16%
• Minimized error in trajectory of a projectile using Mean Shift Clustering to detect hit/miss with confidence map
• Conquered 29th rank out of 170 teams contending in National Robocon 2018
Sign Language Interpreter: June 2017 – October 2017
• Crafted Convolutional Neural Network, trained over images, that translates sign language gesture into text
• Accomplished an accuracy of 63% and precision of 78% after 10 epochs at learning rate of 0.01
• Developed interface system to accept visual inputs in real time for translation
Robocon 2017 August 2016 – March 2017
• Rendered flight trajectory of projectile in footage using Normalized Cut algorithm from OpenCV
• Created tool to control the stroke force of a piston cylinder to correct calculated error
• Achieved 37th rank out of 168 teams performing in National Robocon 2017
Robocon 2016 August 2015 – March 2016
• Devised A* algorithm to navigate out of a maze with record time of 39 seconds for 3-minute task
• Pioneered low power sensors to limit energy consumption for theme of “Clean Energy – Recharging the World”
• Secured 50th position out of 151 teams participating in National Robocon 2016

• Novel use of supervised machine learning to develop an intelligent, real-time, synchronous, dynamic sign language reader for deaf and mute [Patent App No. 201721027128]
• Artificially Intelligent Temperature Monitoring System for Prudent Brooding. [(ISSN 2229-5518) - I0123274]
• A brief survey of finance management [IJSER (ISSN 2229-5518) paper ID: I0118772]

Skills & Technologies

  • Go
  • PHP
  • aws
  • Node.js
  • JavaScript
  • CSS
  • Docker
  • HTML
  • haml
  • Cloud9
  • Java
  • C
  • cpp
  • Ruby
  • python

Project Involvement

Onkar Kulkarni has no publicly viewable Youtube videos.