KRISTINE PETROSYAN

Data Scientist

Austin, TX ***59 Ph: 1-737-***-****

adjhun@r.postjobfree.com github.com/kristinepetrosyan linkedin.com/in/kristinepetrosyan OBJECTIVE: To blaze a career in Data

Science that will enable me to effectively use my

qualifications in Computer Science and Data Science Immersive to develop optimum synergies for the organization. PROFESSIONAL EXPERIENCE:

Data Science Engineer(Personal Projects/Flatiron School) Jan 2020 – Present

Capstone Project: Hotel booking prediction

Predicted whether a booking will be canceled or not to allow the hotel’s revenue manager to make better informed decisions to improve overbooking strategies and cancellation policies.

This project built several classification models to predict each booking’s cancellation outcome. RandomForest Classifier was the best model.

With 88% accuracy, more than eight times out of ten, this model correctly predicts whether a reservation will be successfully booked or cancelled.

The role of booking lead time and country of residence of the presumptive hotel guest in determining the cancellation probability is more pronounced for online than for offline or travel agency bookings.

X-Ray Pneumonia Detection Using Deep Learning

Trained convolutional neural networks of various architectures to diagnose pneumonia using chest x-rays.

Improved accuracy of model from 73% to 83% accuracy using image pre-processing and augmentation techniques, and transfer learning; final model classified 99.7% of all pneumonia cases in the test set.

Utilized BeautifulSoup and Scrapy for parsing through HTML files and Extracting data.

Object detection in TensorFlow TinyML

Trained object detection models specifically for microcontrollers. Used Google Clouds, Jupyter Notebooks and TensorBoard to visualize the EDA.Utilized datashader, pandas, numpy, seaborn, Keras, TenserFlow.

Bank Marketing Deposit subscription prediction

The aim of this project is to predict the result of target variable (subscribing bank deposit) by applying several machine learning classification models.

Used data cleansing, feature engineering, modeling, visualization, Time Series, Image classification modeling.

Seniors at Risk: Using Spatial Analysis to Identify Pharmacy Deserts

Used open source xarray-spatial library to perform analysis on Census Data. Used ML tools in GIS.

Wrote unit tests for testing functionality using pytest to perform the functionality of GIS related functions which only work with xarray-spatial libraries.

COVID hotspots in relation to meat packing locations

Used numpy, pandas, xarray-spatial and Carto libraries to analyze and visualize the results.

Used scikit for building classification models using KMeans.

King County Residential Real Estate Bargain Hunting

Analyzed +20,000 King County home sales to determine how an investor can maximize return-on-investment (ROI).

Built regression models using scikit learn and statsmodels to estimate home sale price. Recommended optimal times of year for buying investment property and King County neighborhoods with upside potential

Python/Django Engineer(freelance at UPWORK) Sept 2015 – Present

Designed the entire infrastructure architecture using AWS services and tools.

Wrote data migrations/ETL scripts using SQLAlchemy and Python for converting data.

Used NGINX and Redis caching to make the site more responsive and scalable.

Utilized Django Rest to implement the rest endpoints exposing business functionality.

Setup Django CMS and developed templates with the help of the frontend designer.

Designed and implemented the data model using PostgreSQL 9.3 in hot standby mode.

Proficient in the use of GIT source version control and GitHub flow for code reviews.

Converted PSD's into HTML Django templates using HTML5, CSS and JavaScript.

Wrote Ansible configuration management scripts for provisioning systems. Junior Python Developer(VTG Armenia) Nov 2012 – Apr 2015

Gathered business requirements and translated them into functional and technical specifications.

Used Python 3.0 for generating IP access frequency lists in different data logs.

Wrote Python scripts to read from Excel files and generate XML configuration files.

Worked under the agile methodology in a sprint of 2 weeks and monthly releases.

Used Django 1.5 as the web application framework to maintain the customer facing website.

Implemented the data model using MySQL as the relational database system. TECHNICAL SKILLS:

Machine Learning: classification, regression, clustering, feature engineering

Statistical Methods: time series, regression models, hypothesis testing and confidence intervals, principal component analysis and dimensionality reduction, stochastic differential equations (SDEs)

RDBMS: MySQL, PostgreSQL, MongoDB

Software and Programming Languages: Python(Scikit-learn, Numpy, Scipy, Pandas, Keras, TensorFlow), Java, C ++, Hadoop (Hive, MapReduce), Linux, Oracle, Microsoft Excel

Operating Systems: Windows, Ubuntu, MAC

Web Technologies: XML, JSP, Servlets, JavaScript, HTML5,CSS, Scrapy

App /Web Servers: Tomcat, JBOSS

Versioning Tools: Subversion, GIT

EDUCATION:

Flatiron School, New York, NY

Certificate in Data Science, 2020

American University of Armenia (University of California), Yerevan, Armenia Master in Business Administration (M.B.A) in Finance, 2014

State Engineering University of Armenia, Yerevan, Armenia B.E. in Computer Science, 2009

Contact this candidate