LEADERSHIP BEYOND THE CLASSROOM
DATA SCIENCE SOCIETY @ BERKELEY
Contributed to the Google Chatbot Metrics Project and the Academic Development Committee at DSSB.
Academic Development Committee:
Developed workshops related to R, Python, SQL, Excel and RapidMiner to promote Data Science Awareness on campus. Here is an image of me, giving a lecture about SQL at a workshop aimed at Berkeley Haas students (hosted by Data Science Society @ Berkeley and Haas Business School Association). At the end of the workshop, my teaching received an average rating of 4.65/ 5 by a group of 60+ students. You can view the workshop’s lecture slides here.
Google Chatbot Metrics Project:
Developed linear regression and classification models to generate metrics that evaluate chatbot versions by applying machine learning. Compiled the query/response database for different chatbots (Siri, YokoBot, Cortana etc) and then created a Document Term Matrix for initial version of YOKO bot and again for the manually created version of YOKO bot. Extracted features, concatenated these two datasets together with appropriate 0 and 1 labels, built a machine learning model to determine an AUC (Area Under the Curve) score and rescaled to 0-1 range.
Project Background:
A chatbot is a computer program designed to simulate conversation. Over time, chatbots have become more sophisticated. Call the earliest version of a chatbot Version 0 (V0). Call the next version of that chatbot V1. And so on. The difference between Vn and Vn+1 could be various things, including, but not limited to, additional features, bug fixes, personality, and knowledge. The goal of this project is to construct a function that maps Vn and Vn+1 to a number that quantifies their differences. Was there a minor bug fix or did the new version add tons of new capabilities? Such a function will contribute to metrics at Google.
MICROSOFT ENTERPRISE CLOUD & MOBILITY HACKATHON
Our team won the Grand Champion trophy for this hackathon. Contributed to an optimization project, developed using Azure Machine Learning, that would potentially save over a million dollars annually in operational expenses for Microsoft's customers. Helped the team decide the regression models to predict the successive concurrent sessions, which affected the number the virtual machines required for hosting the active users. This project also bagged the "People's Choice", "Best Hack" and "The Guru - Best Learner" awards.
CALHACKS 2.0
Won the Second Runner Up Award of $500 for Best Use of a Magnet API at CalHacks 2015, UC Berkeley's annual intercollegiate hackathon. Developed an Android app (called “Politically Correct”) with 4 other teammates, to prevent users from sharing offensive content. The app uses the Magnet Messaging API to send your content anonymously to 10 people chosen at random who upvote it if it’s not offensive and downvote it if it’s offensive. It also uses a logistic regression model deployed on the Microsoft Azure Machine Learning service to automatically classify content as offensive and non-offensive. Based on both results, users can choose whether or not to share their content. This project was implemented in Java and Python.
Project Link- http://devpost.com/software/politically-correct
MACHINE LEARNING @ BERKELEY
Contributed as a Software Developer to ML@B|X (Machine Learning @ Berkeley) in Spring 2016. I contributed by developing an Android app which formed the user interface for the Autonomous Robot Tailor Project. I learned about linear regression and Principal Component Analysis by working on this project. You can read more about the club and its projects here.
Project HOPE
I was a part of IEEE Project HOPE (Hands on Practical Electronics) in Fall 2014. Through this project I got practical experience and exposure to Texas Instruments laboratory equipment at UC Berkeley and I supported the promotion of technical involvement among non-STEM majors.
GAMESCRAFTERS
In Fall 2014, I was a part of a UC Berkeley Research and Development group called GamesCrafters, under the guidance of Dr. Dan Garcia. This project involved solving, analyzing, programming and exploring the fertile area of combinatorial and computational game theory. At the core of the project was "Gamesman", a system developed for solving, playing and analyzing two-person, abstract strategy game.
IEEE Industrial Relations Team
In Spring 2015, I was a member of the organizing committee to host technical infosessions, workshops and startup fairs. My role was to contribute in strengthening IEEE’s industrial ties and relations with various student groups on-campus.
