Hey! I am

Megala Kannan

I'm a

About

About Me

I am a Data Scientist at PepsiCo with a graduate degree from Columbia University, Data Science Institute. I have obtained my undergraduate in Computer Science from India. As a part of my graduate program I have acquired various skills to become a full fledged Data Scientist. I have studied courses like statistical modeling, machine learning and applied deep learning.

My main interests lies in applied machine learning and deep learning techniques to solve real world problems. I enjoy working in natural language processing domain as well. My course projects have enabled me to gain the required skills to be a well rounded data scientist. I worked at Johnson & Jonhson as a data science intern where I sparked an interest in the healthcare domain. The problems being worked on in healthcare are very challenging and gave me a good exposure to real world data science applications.

Apart from aspiring to be a top notch data scientist I enjoy theatre, music and art and hope to someday walk the stage again and get involved in performing arts.

0 Project complete

Download CV

Education

2018-2020

Master of Science in Data Science

Columbia University

Courses: Machine Learning for Data Science, Algorithms in Data Science, Probability and Statistics, Exploratory Data Analysis & Visualizations, Statistical Inference and Modelling. Applied ML, Personalization theory & application, Applied Deep Learning, Computer Systems for Big Data

2014-2018

Bachelor of Science in Computer Science

College of Engineering, Guindy

Courses: Algorithms, Data Structures, Operating Systems, Computer Architecture, Database management systems, Artificial Intelligence, Compiler Design, Theory of Computing, Software Engineering

Jun-Jul 2017

Foreign Technical Training Program

National Chin-Yi University of Technology

Summer exchange program on Computer Science and Electronics

Experience

Apr '20 - Present

Data Scientist

PepsiCo

I am a Data Scientist in the Data Science and Analytics team at PepsiCo eCommerce. I primarily work on using data science techniques to optimize sales and ROI for products sold by PepsiCo

Jun-Jul 2019

Data Science Intern

Johnson & Johnson

Analyzed voice of customer data regarding drug products using Natural Language Processing. Developed medical ontologies using Linguamatics and word embeddings (Fasttext) techniques to perform semantic querying. Built a text analyzer for the voice of customer data using unsupervised clustering models in python.

May-Jun 2017

Data Science Intern

PurpleSlate

Modeled the users’ real time data using clustering techniques. Analyzed price movements in the financial and customer markets.

Skills

Machine Learning

90%

Deep Learning

80%

Statistical modeling

90%

Linear Algebra

75%

Databases & OS

85%

Big Data Technologies

70%

Natural Language Processing

85%

Data Visualization

80%

Tools & Languages

Python, R, SQL, TensorFlow, Keras, Javascript

Numpy, Pandas, Scikitlearn, Matplotlib, Nltk, PySpark, D3.js

Awards

2017

Government of Tamil Nadu award under rule 110

College of Engineering, Guindy

Foreign Technical Training Program

2015-2016

Bicentenary Engineering College Co-operative Society Common Good Fund Endowment

College of Engineering, Guindy

Best outgoing student in BE Computer Science and Engineering

2015-2016

Samuel Memorial Prize

College of Engineering, Guindy

Proficiency in Engineering Graphics

2014-2015

Bicentenary Engineering College Co-operative Society Common Good Fund Endowment

College of Engineering, Guindy

Highest total marks in BE Computer Science and Engineering

Publication

2017

Emotion based music player for Android

College of Engineering, Guindy

Imparted a machine learning approach to perform facial emotion analysis and digital signal processing on audio signals. Categorized songs into various emotions by extracting midterm features to compute their valence and arousal values. Applied regression through Support Vector Machines was used to train the model on these audio features and a Valence-Arousal coordinate plane was defined to segregate the emotions.

Publication Link

Projects

Jan - May '19

News Data Analytics

Microsoft Research

Created a platform to analyze news articles and events. Topic matching of the news articles with Wikipedia pages. Implemented Wikipedia category search tree to obtain categorization of articles.

Github Repo
Sep-Dec '19

Detecting Cancer Metastases in gigapixel pathology images

Applied Deep Learning, Columbia University

Detected tumor cells from pathology images using image segmentation and classification. Constructed a convolutional neural network model using tensorflow and keras frameworks. Designed evaluation metrics to diagnose the presence of cancer in the cells.

Github Repo
Oct-Dec '19

Recommendation system on yelp dataset

Personalization Theory, Columbia University

Built production grade recommendation systems on the yelp dataset for various businesses. Predicted ratings on active users using collaborative filtering, non-negative matrix factorization. Designed a ‘wide and deep’ learning model for user recommendation.

Github Repo
Sep-Oct '19

Movie recommender system

Personalization theory, Columbia University

Built production grade recommendation systems on the movie lens dataset. Designed user-based collaborative filtering and model-based matrix factorization using PySpark ML methods Predicted users top 10 movie recommendations and developed evaluation metrics for the recommender model.

Github Repo
Sep-Dec '18

The global refugee crisis

Exploratory data analysis and visualization, Columbia University

Identified the humanitarian crisis using UNHCR population of concerns data. Explored the datasets to obtain the countries with highest refugee population using R tidyverse. Visualized the flow of refugees across the years using D3.js

Github Repo
Jun-Dec '17

Context based hashtag recommendation system

Software development lab, College of Engineering Guindy

Performed topic modelling on twitter posts using Latent Dirichlet Allocation. Evaluated the results using topic coherence to generate most relevant hashtags. Further analysis on subtweets was performed to obtain better granularity of results.

Interests

My Interests

Things that interest me that are NOT in the data science space!

Extra Curricular

During college I was a part of a theatre club called Theatron where I was involved in several productions for MYTF and CTI by Crea-Shakthi. I love acting and being a part of the production team for dramas. I organized theatre events at culturals like street play and mono acting. I was the head of finance in my senior year at college. Outside of college I also worked for EVAM and I was a part of the organising team for The Hindu Theatre Festival in 2016 and 2017.

Volunteer

In my college days I was an active part of the Youth Red Cross foundation and Rotaract Committee that organized events for college students to help for the well being of the society.

Hobbies

I am a trinity level 6 pianist and I enjoy playing some of my favorite songs in my free time. I also like to work on some DIY arts & crafts projects for room decor or to gift to my friends as presents. When I am trying to have a lazy day I love to binge watch youtube videos of some of the lifestyle vloggers I follow and occasionally I like to shoot some vlogs myself

Contact

Contact Me

Address

New York, NY 10027

Contact Number

+1 908-693-3290

Email Address

msk2245@columbia.edu