mobile_bg

About me

Eleven years ago, I developed my first computer program, a basic calculator, in C# and I was hooked right away. Over the years, I worked as a Software Engineer for two years and published an iOS app. My undergraduate Statistics minor changed the way I think about data and since then, I have enjoyed analyzing how my personal data (such as sleep, nutrition, and running pace) interact with one another. I started learning about Machine Learning soon after as it gave me a unique opportunity to apply my statistics knowledge in combination with my software engineering skills. To learn more about this field, I studied Machine Learning in the Master of Engineering program at UC Berkeley.

During my summer 2019 internship at an early-stage startup, I learned about and used new techniques of modeling and data collection to build on my existing machine learning knowledge. My tasks involved training image classification models with TensorFlow to predict a hospital bed’s patient-occupancy score and collecting training data by scraping images and converting YouTube videos using a video annotation tool that I built with OpenCV.

At UC Berkeley, my team researched traffic light coordination using deep reinforcement learning with an open-source traffic control framework called Flow. Our work, in improving Flow over the past year, has allowed me to apply the new skills that I learned in technical courses such as Machine Learning Systems and Deep Reinforcement Learning.

In 2017, I published an iOS app called Drinks Tracker on the App Store. I created an algorithm that calculates the user’s BAC level based on multiple metrics, which the app used to notify the user about their estimated Blood Alcohol Content (BAC) level. The app was installed by over 3.9k users and got 104k impressions when Apple removed this category of apps from the App Store.

My current fitness goal is to complete a 7.5 mile run in an hour. Currently, I am two minutes away from making it!

In my spare time, I enjoy watching movies and TV, especially science fiction classics such as Star Trek and Doctor Who. 🖖

Currently, I want to work as a Machine Learning Engineer, in the Bay Area, to make better use of all the data that we generate every day.

Researched traffic light coordination using Deep Reinforcement Learning

with Prof. Alexandre Bayen

Fall 2019

Worked in a team to improve an open-source traffic control framework called Flow

Updated Flow to use AIMSUN’s (traffic simulation software) improved API methods

Trained with PPO and real-world peak-hour traffic data from Arcadia, CA

Optimized the time offsets of five arterial traffic lights to increase average traffic speed
Artificial Intelligence to improve City-scale Traffic Flow

Researched how to reduce mixed-autonomy traffic congestion using Deep Reinforcement Learning

with my Capstone team and Prof. Alexandre Bayen

2018 – 2019

Optimized the behavior of autonomous vehicles to increase average traffic speed in a city-scale scenario

Worked in a team to use and improve an open-source traffic control framework called Flow that simulates traffic with SUMO and trains with Ray
Drinks Tracker

Published an iPhone app on the iOS App Store

Drinks Tracker

2017 – 2018

In 2017, I published an iOS app called Drinks Tracker on the App Store. I created an algorithm that calculates the user’s BAC level based on multiple metrics, which the app used to notify the user about their estimated Blood Alcohol Content (BAC) level. The app was installed by over 3.9k users and got 104k impressions when Apple removed this category of apps from the App Store in 2018.

GitHub
This project was uploaded to GitHub in 2019.

Completed the Machine Learning by Stanford course

Taught by Andrew Ng, Co-founder, Coursera and Adjunct Professor, Stanford University

Spring 2017

Meeting assistant

Created a voice-based meeting assistant as our Capstone team project

for PureCloud by Interactive Intelligence

Spring 2016, The Ohio State University

Analyzed speech during video calls to automatically create meetings in Google Calendar

Worked with the IBM BlueMix, Google Cloud Speech, and Google Calendar APIs

Analyzed and summarized the text of a diverse collection of Reuters articles in Python

Spring 2016, The Ohio State University

Used Beautiful Soup 4 and scikit-learn to implement several clustering and classification algorithms such as DBSCAN, Hierarchical Clustering, Naïve-Bayesian, and K-Nearest Neighbors

Calculated Jaccard Similarity scores and Mean Squared Error to accurately create association rules

Developed several Android apps as part of coursework

integrated with the Twitter, Google Calendar, and Google SafeSearch APIs

2015 – 2016, The Ohio State University

'Best Lifesaving Hack’'

Won Best Lifesaving Hack at Qualcomm’s intern hackathon for an Augmented Reality app

Summer 2015, Qualcomm

Created an Android app to visualize sound to assist the hearing impaired
isaosu.com

Organized philanthropic and cultural events for 300+ people

as Webmaster and Chief of Staff (Indian Students Association)

2013 – 2016, The Ohio State University

uncakd.org

Aaj Ka Dhamaka

Webmaster

Fall 2013, University of North Carolina

Advanced Energy Vehicle project

Model Monorail

Spring 2013, The Ohio State University

Machine Learning Engineering Intern

LookDeep Health

Summer 2019, Oakland, CA

Trained image classification models to predict a hospital bed’s patient-occupancy score

Built my model’s training pipeline in TensorFlow 2 and learned to use DVC, an open-source model versioning system

Collected training data by scraping images and converting YouTube videos using a video annotation tool that I built with OpenCV. Also used and contributed to Scalabel, an open-source annotation tool

Generated 40k randomized synthetic training images after learning to use Blender

Improved my model using computer vision techniques with transfer learning on InceptionV3 and ResNet-152

Assisted another intern in developing the end-to-end system from reading images off the camera to serving predictions using TensorFlow Serving and displaying the output inference in the browser

Software Engineer (Full Stack)

OCLC (Online Computer Library Center)

2016 – 2018, Columbus, OH

Wrote Java code while updating 7 legacy webapps, used by 15k+ administrators for library management (books, media, etc.), to use jQuery, PostgreSQL, and Apache Tomcat

Enhanced OAuth authentication services used by OCLC’s enterprise partners to support mobile apps

Supported a $3M+ contract by writing utilities to easily internationalize a webapp in 20+ languages

Optimized database handling and functionality for compliance with GDPR requirements

Built and deployed our webapps to several QA and Production environments

Interim Engineering Intern

Interim Engineering Intern

Summer 2015, San Diego, CA

Improved my internal webapp from 2014 and created new features to automate tasks for other team members

Automated tasks using C# to internalize SQL scripts into an ASP.NET application

Interim Technology Intern

Qualcomm

Summer 2014, San Diego, CA

Overhauled an internal webapp used to manage a portal which simplified doctor-patient communication

Resolved and implemented requested user stories to increase operational productivity

Made extensive use of ASP.NET, SQL, HTML5, and JavaScript while collaborating with other team members using Team Foundation Server in Visual Studio

The Ohio State University

CSE Help Desk Operator

2014 – 2016, Columbus, OH

Assisted Computer Science professors, faculty, and students with their technical queries and issues

Replied to any queries sent to help@cse.ohio-state.edu diligently and in a timely manner

Printed posters for graduate students

Ensured that all the CSE printers (~20) are working at all times with no paper and ink issues

ICICI Prudential AMC Ltd.

IT Intern

Summer 2010, Mumbai, India

Gained experience programming in work environment

Primarily programmed in ASP.NET and C#

Created websites for inputting and recalling user details

Worked in Microsoft Access to create databases to store user details
Drinks Tracker

Drinks Tracker

In 2017, I published an iOS app called Drinks Tracker on the App Store. I created an algorithm that calculates the user’s BAC level based on multiple metrics, which the app used to notify the user about their estimated Blood Alcohol Content (BAC) level.

The app was installed by over 3.9k users and got 104k impressions when Apple removed this category of apps from the App Store in 2018.

GitHub

This project was uploaded to GitHub in 2019.

App Store Description

Drinks Tracker allows you to search a list of nearly 4,000 drinks and easily add drinks to track your Blood Alcohol Content (BAC) percentage.

Drinks Tracker Drinks Tracker Drinks Tracker Drinks Tracker Drinks Tracker Drinks Tracker Drinks Tracker Drinks Tracker Drinks Tracker Drinks Tracker

Features

  • Search and easily add drinks. Favorite your frequent drinks for easy access!
  • Track your BAC using the Today View widget and the Apple Watch app. Complications and Dynamic Notifications are coming soon!
  • Easily add custom drinks and favorite them in a few taps.
  • Get notifications when your BAC is at 0.08% and 0% and a reminder to add your next drink.
  • Customize your profile to get an estimated BAC tailored to you!
  • 3D Touch the app icon to quickly add a drink and to see your BAC in the widget.
  • Suggest drinks so that they can be added to the list!