Sarthak Khanna

Salt LAke City, Utah, US
khannasarthak.1994@gmail.com

I'm a software engineer interested in web technologies, machine learning, and data sciences, I also have a background in electronics engineering.

View a PDF version of my resume here.

Experience

Backend Software Engineer

SimpleCitizen

Designed and created new software features for the single page application using Python3, Django, GraphQL, PostgreSQL and Rest API's.

Modified existing database models to increase performance efficiency of CRUD processes that had complex cascading designs.

Implemented Single Sign On and Multi-Factor authentication using SAML for the enterprize side of the business.

November 2019 - Present

Junior Python Developer

MarketDial

Designed, implemented and tested ETL pipelines to extract data from client source systems (Google Cloud Storage buckets), cleaning them to be loaded into our Postgres and BigQuery databases using Python and SQL.

Created and documented a new feature giving clients the ability to A/B test based on their customer demographics along with product categories using Python and SQL.

Overhauled existing ETL code to use BigQuery instead of Pandas (python). Resulting in a drastic decrease in data processing time and increased performance by reducing system resource usage.

Made various optimizations and automated other ETL processes making them more streamlined and easier to use, leading to better back end experience for developers.

January 2019 - November 2019

Data Analytics Teaching Assistant

Trilogy Education

Mentor participants of the Data Analytics Bootcamp at UNC Charlotte with various technologies like Python, JavaScript, HTML/CSS, Machine Learning and also data visualization and dashboarding tools like D3.js, Tableau, etc.

Support the instructor of the ongoing class with content delivery and grading of homework and projects.

Aid participants in reaching technical milestones thus inculcating a developer mindset in them.

November 2018 - January 2019

Technical Developer ( Part Time )

Stealth Gaming

A social gaming start-up, developed the website and implemented a web based user maintenance system.

Quoted and procured computer infrastructure for setting up gaming systems for the lounge, also maintained the technical infrastructure.

Developed marketing solutions to increase revenue and customer loyalty.

May 2014 - May 2016

Education

Arizona State University ( ASU )

Master of Science (MS)

Computer Engineering (Computer Systems)

June 2016 - May 2018

Vellore Institute of Technology ( VIT )

Bachelor of Technology (B. Tech)

Electronics and Communications Engineering

June 2012 - May 2016

Skills

Programming Languages

Python3
PostgresSQL
JavaScript
Hylang
Kotlin
HTML5
CSS3

Web Technologies

Django, GraphQL, REST API's, Flask, Bootstrap, jQuery, JSON, MongoDB

Frameworks/Libraries

PyTest, PyParse, scikit-learn, numpy, Pandas, matplotlib, faker

Tools, OS & Misc

Git, Linux, Windows, PyCharm, Google Cloud Services, AWS, Agile (Scrum)

Projects

BraiNet: Unlock phone using EEG signals

An android app to authenticate users based on their brain EEG signals.

Setup cloud and fog server.
Implemented the SVM algorithm to train and compare the users to authenticate.
Created an offloading algorithm to determine which server to choose for authenticating.
Measured various metrics of performance. Detailed information in the code repository.

Airbnb New User Bookings Prediction

The project predicts the first booking that a completely new user will make upon visiting airbnb’s website. (Kaggle problem statement)

Processed raw data, making it usable for further analysis using the python pandas library.
Performed feature engineering and statistical analysis of the data obtaining relevant insights using scikit-learn.
Implemented traditional classifiers and observed the accuracy. Tested blended ensemble of classifiers using multiple data sets.
Analyzed blended ensembles’ performance versus traditional classifiers.

Berkely CS188 Pacman AI Challenges

The project predicts the first booking that a completely new user will make upon visiting airbnb’s website. (Kaggle problem statement)

Search algorithms: A star, Uniform Cost, DFS, BFS.
Programmed Expectimax, Minimax and alpha beta pruning.
Value Iteration, Q - Learning and policies.

Activity Determining Mobile App

An android app to determine whether the user jumped, walked or ran using machine learning.

Created the android app using android studio.
Used SVM to train on the data and create a model with 30 instances of each activity.
The K-fold (k=4) cross validation gave an accuracy of 94.805%

Theater Management system

A project to manage movie screenings, performances, bookings and customer data for a multiplex.

Built the web application from scratch using python flask and MySQL.
The database had a relational model with queries handling tasks like insertions, updates, deletions. Triggers and indexes were also used.
Generated the dummy data using python faker library.

Real Time Accelerometer Plotting

An android app to plot users real time accelerometer data, uploads to a server and then download when needed.

App plots and stores the data for every user.
It also uploads the data to a web server, which can be downloaded later if needed. IMplements a manual sync function of sorts.

Naïve Bayes, K-NN, K-means implementation

Implementation of the said algorithms on standard data sets.

Prepped raw data and realized the algorithms in python. Used numpy and matplotlib libraries.

Stuttered Speech Recognition

B.Tech final year project, designed algorithm to remove stuttering from a voice sample and passed to Google voice API for recognition.

Modified an existing algorithm to facilitate speech recognition for people suffering from stuttering.
Performed audio processing and stutter removal in Matlab.
Designed a neural network to learn from training samples to remove stutter from new test inputs in python, this new audio was then passed to the Google speech API to be recognized without the stutter.

Digit recognition using Matlab

Created a neural network to recognize hand written digits.

Programmed a backward propagation neural network to recognize hand written digits using Matlab.
Used the MNIST database to train the neural net.

Polygone - Internet of Things

An idea based on providing real time data analysis to aid user decision making used IoT and connected smart devices.

Created a dummy web app to showcase the possibilities of the idea.
The project was awarded the second place in the hackathon it was formulated.

Web Development: RNR

Created a website for a local business (gaming lounge called Rock n Rollas).

Used HTML, CSS , Jquery and BootStrap Responsive frame work.
Set up domains, servers, facebook pages.

Home Automation System

This concept is a modular home automation system with available add ons.

The project consisted of making a prototype as proof of concept, a temperature controller was made, where the user could regulate the temperature using a mobile app.
Arduino, Xigbee, NFC and WiFi modules were used to create the network and MIT app inventor to create the prototype mobile app.
This was made at a Honeywell Hackathon, this idea was selected among the top 10 and we were given a chance to go to Honeywell Bangalore to meet prospective VC’s and showcase our creation.

Game Development using Unreal Engine 4

A LAN multi player 3rd person shooter game.

Worked on the back end to create the LAN capabilities and also made the various logics of the game ( like HP, teleportation and special moves).

Game Development using Unreal Development Kit

The project was a third person puzzler game made on unreal development kit.

Made two levels of a 3-D puzzler game.
Designed and constructed the levels along with game logics and back end scripting using UnrealScript.

Propeller Clock

Based on the principle of persistence of vision, the project displays an analog clock on an array of rotating LED’s.

Created the propeller which has an array of LED’s, which give the illusion of an actual clock.
Coded the IR sensor to detect RPM and change speed accordingly using an Arduino.

Notifier and Connectifier

Prototyped a cost effective Google glass alternative.

Used the center piece of a Sony smart watch, basic principles of reflection and refraction.
The frame was made using PVC pipes connected on top of a traditional spectacle frame.
Capacitive buttons on the the smart watch were extended using conductive fiber to give the user easy control.
The glass can be paired with any bluetooth enabled smart phone.
The project was created in under 24 hours at a college hackathon.

Baby Vitals Monitoring

Created low-cost sensor network to monitor neonatal vital signs.

Developed an ingenious heart rate sensor using a plastic bottle cap and a ball point pen spring.
Used temperature sensor and developed a heart rate monitor as well.
All these were connected using bluetooth using Zigbee and utilizing the power of the Arduino board.
The readings were visible on a simple mobile app.
This device was made at a hackathon organized by CAMTech.