Harsha Wardhan Reddy Duvvrv
Santa Clara, CA *****
firstname.lastname@example.org, 669-***-****,www.linkedin.com/in/harshawardhanreddyduvvrv EDUCATION
Master’s in Business Analytics/Data Analytics, Santa Clara University, CA, US December 2019 Bachelor of Engineering, Visvesvaraya Technological University, Bangalore, INDIA June 2016 TECHNICAL SKILLS
• Programming Languages & Tools: Python(NumPy, Pandas, Scikitlearn, SciPy, OOPs), Scala, Linux, R, C++, Flask, SAP ABAP, SAP XI, Keras, TensorFlow, MS Office, Agile.
• Database & Modelling Tools: MYSQL, PostgreSQL, Navicat, Erwin, Data Warehouse(Star Schema, Snowflake Schema), Pentaho.
• Data Science: Machine Learning, Deep Learning(CNN, RNN), Reinforcement Learning, Supervised Learning(Linear Regression, Logistic, Decision Tree, Random Forest, Naïve Bayes, GLM, KNN), Natural Language Processing, Regularization(Lasso, Ridge), A/B testing, AdaBoost, Bagging, Marketing Mix Modelling, ARIMA, Monte Carlo, Clustering, Semi-Supervised Learning.
• Big Data: MapReduce, Hadoop, HDFS, Hive, PySpark(RDD, DataFrames), SparkSQL.
• Visualization: Python(Matplotlib, Seaborn), R(Shiny), Tableau, PowerBI, Plotly-Dash.
• Cloud: Azure ML studio, Event hub, Stream Analytics, Azure SQL, HDInsight, Databricks, Azure Storage, AWS Redshift, S3, EC2. EXPERIENCE
Data Analyst Intern – Nisum, Fremont, US June 2019 – Sep 2019
• Built an intelligent and scalable personalized recommendation system using machine learning models (SVD, clustering, Random forest). The model deals with the cold start problem and efficiently recommends products based on the customer's unique interests and preferences. Presented the model to product owners/clients.
• Built a data pipeline in Microsoft Azure to do real-time analytics(recommendation) which involves ingesting, storing, visualizing, cleaning customer ratings, product details, and customer details dataset. Eventus Systems - US Jan 2019 – Dec 2019
• Worked with Eventus Systems to come up with a new approach to identify market manipulation by the traders.
• Cleaned and organized the market data, clients trading data into a PostgreSQL database to further perform feature engineering, visualization, and analysis using unsupervised learning.
• Results of using asymmetry features on client activity and measuring their market impact correctly identify periods of unusual trading activity.
Software Engineer - Accenture, Bangalore, India 2016 – 2018
• Communicated with client and third party to understand the functional requirement to calculate balance from the account expense data and to calculate sales for a specific time period.
• Worked in a team to develop a program to extract account expense data along with the sales data from the database, generated excel files to provide account status and sales information to the client via a job scheduler.
• Collaborated on a team of 5 to add custom KPIs for Finance and Procurement functionalities by applying object-oriented ABAP. These KPI’s were vital for the client to evaluate the performance of the business.
• Developed custom reports automating DML operations on different tables. ACADEMIC PROJECTS
Buy Online Pick Up In-Store (BOPS)
• Utilized the past 3 year’s data of national jewelry retailer to analyze the impact of BOPS on online channel sales and returns. Performed data cleaning, aggregation, visualization, and derived managerial insights.
• Implementing BOPS decreased online channel sales by 49% and returns by 66%. BOPS strategy reduced sales and returns on product-level by 30% and the sales value of low-priced items dropped. Ticket Booking system
• Designed and implemented OLTP, Data Warehouse in MySQL, using Pentaho to do ETL for the movie ticket booking system.
• Implemented star schema by creating fact tables, dimension tables, Slowly Changing Dimension.
• Further connected Tableau with the Data Warehouse to provide managerial level insights by generating descriptive and predictive dashboards.
Credit Card Fraud
• Analyzed credit card transactions dataset using various supervised learning algorithms to identify fraudulent transactions.
• Employed various methods to deal with an imbalanced dataset as well as to improve the AUC score. Best Buy 1 Month Sales Examination.
• Identified potential customers having a high propensity to purchase Best Buy’s warranty Plan applying the Logit and Probit model.
• Interpreted results are helpful for customer segment-based advertisement strategy.