Data Software Developer

Gaurang Vilas Gatlewar

Address: **, ****** *, ********, ** 11223 Email: ******@***.*** Tel: +1-347-***-**** LinkedIn: https://in.linkedin.com/in/gaurang-gatlewar Github: https://github.com/GaurangGatlewar HackerRank: https://www.hackerrank.com/Gaurang_Gatlewar Education: Computer Science

New York University (2018)

Masters of Science[3.5/4]

Indian Institute of Technology (2016)

Bachelor of Technology[3.5/4]

Technical Skills:

Languages: C, C++, Java, Python

Web Technology: HTML, MySQL, PHP,

CSS, JavaScript, Flask.

Tools: MS Office, Amazon AWS,

MATLAB, AutoCAD, CATIA, Wireshark,

Jupyter Notebook, VirtualBox, GitHub,

Metasploit, Hadoop, MapReduce, Pig,

Spark, Docker, Tensorflow.

OS: Windows, Linux, Mac.

Coursework:

New York University:

Machine Learning

Deep Learning

Introduction to Data Science

Programming for Big Data Analytics

Design and Analysis of Algorithms

Data Structures and Algorithms

Penetration Testing

Network Security

Computer Networking

Introduction to Operating Systems

Computer Architecture

Coursera:

Python Data Structures

Using Python to access Web Data

Using Databases with Python

Cryptography 1

Algorithms, Part 1

Algorithms, Part 2

Neural Network and Deep Learning

Machine Learning with Big Data

Extracurricular Activities:

2400th in India in IIT-JEE (99

percentile).

Ranked 606th (96.48 percentile) in

IMO (International Mathematics

Olympiad).

Event organizer, NYU Center for

Cybersecurity.

Selected by Goldman Sachs as a

Quantitative Analyst.

Departmental Representative in

Department of Undergraduate

Committee.

Ranked 214/2453(Gold) in

NCL CTF 2018.

Coordinated with Kiran Society (NGO)

and improved the KAFO (leg caliper)

design.

EXPERIENCE

Revmax Corp, New York, USA

Software Developer Intern May 2017-July 2017

● Scraped 7+ websites using BeautifulSoup library and extracted json data from 15+ rest APIs to gather 25+ GBs of data spanning 30 years.

● Performed ETL by setting up a cron job using Amazon AWS Lambda function using a python script along with its dependencies in a virtual environment resulting in an automated data collection process.

● Co-ordinated with a team to setup a ML model to predict the probability of hiring a cab and achieved an 87% cross-validated accuracy in 30% of the area which was determined by setting up an 80% confidence level threshold. Larsen & Toubro Infotech, India

Research Summer Intern May 2013-July 2013

● Coded in MATLAB and C/C++ to design Whitcomb Rule compliant UAV fuselage reducing drag thus leading to supersonic flight capability.

● Modified design to include delta wings with NACA airfoil providing better gliding capacity increasing the flight duration by 30%.

● Simulated the flight in MATLAB flight simulator and optimized parameters to maximize the efficiency.

ACADEMIC PROJECTS

Taxi Demand Prediction:

Technologies Used: Python, PySpark, MongoDB, Web Scraping, Hadoop, R, Django, Rest APIs, Data Analytics, Machine Learning, MapReduce, Pig, Cluster Analysis.

● Performed ETL to gather data(>10GB) from 3 data sources.

● Cleaned, modified and merged the data into a single dataset for prediction model also making it scalable for Big Data.

● Predicted demand using Random Forest Classifier achieving a 90% accuracy.

● Optimized the model using PCA and GridSearchCV increasing the accuracy to 93% cross-validated accuracy.

Twitter Data Analysis:

Technologies Used: Python, Twitter API, Machine Learning, Jupyter Notebooks, Numpy, Pandas.

● Identified twitter “Bot” profiles using Machine Learning Classification algorithms based on the attributes like name, location, verified, etc.

● Extracted tweets using Twitter Rest API and processed them using Python libraries such as pandas and numpy in Jupyter Notebook.

● Tested multiple algorithms and determined the best suited algorithm.

● Optimized the performance using GridSearchCV to predetermine the optimal input parameters which increased the probability by almost 2%.

● Ranked 1st in Kaggle competition by achieving >95% accuracy. Artificial Neural Network based Galling frequency estimation: Technologies Used: Matlab Neural Network Toolbox, C/C++.

● Gathered, Validated and Tested data against 12 algorithms for best estimation and chose Levenberg-Marquardt and Bayesian Regularization algorithms based on the minimum error criteria using MATLAB Neural Network Toolbox in C programming language.

● Tested multiple Neural Networks and finalized a 3 layered neural network and validated the results with available data to generate a Neural Network capable of >90% accuracy.

Google Landmark Recognition Challenge:

Technologies Used: Python, Tensorflow, NYU HPC.

● Created a dynamic batch processing python code to manage 400 GB data.

● Created a deep learning CNN to classify landmarks in 15k categories.

Contact this candidate