Sales Sql Server

Karan Kanchan

** ******** ***, ******, **, 021**-***-*** 1764 *******.*@*****.***.***

EDUCATION

Northeastern University, Boston, MA Dec 2017

Master of Science in Engineering Management

University of Mumbai, India. May 2014

Bachelor of Mechanical Engineering

SKILLS

Time Series Analysis, Deep Learning, NLP, Computer Vision, Recommender Systems, Reinforcement Learning, Operations Research, Project Management, A/B testing, Social Media Analysis. Programming Languages: R, Python, Matlab, Java.

Libraries: sparklyr, numpy, scipy, pandas, scikit-learn, tensorflow, keras, pytorch. Databases: MySQL, SQL Server, Oracle 11g Express, PostgreSQL, mongoDB. Big Data: Apache Hive, Apache Spark, SparkSQL, Spark Streaming, MLib, GraphFrames Data Integration: Talend Enterprise Data Integration, SQL Server Integration Services (SSIS). Business Intelligence: Tableau, Qlik Sense, Microsoft Power BI, RStudio Shiny. Tools: Azure ML, AWS, SSDT, Visual Studio, Orange Data Mining, Jupyter / IPython notebook, Toad. RELEVANT WORK EXPERIENCE

Charles River Laboratories Inc., Data Scientist Co-op, Wilmington, MA January 2017 – August 2017

Conducted exploratory statistical analysis and designed time-series; ARIMA, Holt Winters, ETS as well as regression algorithms to predict revenue/sales 12 months into the future with accuracy up to 91.07%.

Designed Long Short Term Memory (LSTM) Deep Learning models using Recurrent Neural Networks

(RNN) to increase prediction accuracy for sales and revenue by 4-5%.

Coded production ready R, Python and SQL scripts to automate the process of fetching data in R environment, building Artificial Intelligence models, picking the best one, and pushing predictions back to SQL Server database for over 900 scenarios.

Reduced training time by 70% and generated multiple Power Bi reports based on the analysis.

Mined HR data to predict attrition using Decision Trees, Bayesian techniques and Logistic Regression. APPLIED PROJECTS

Computer Vision

Developed algorithms to detect cars, buses, pedestrians, etc. using You Only Look Once model.

Built a face recognition system which can both recognize and verify a person.

Implemented neural style transfer algorithm to generate novel artistic images. Deep Learning

Built a multi-class Deep ANN image classifier to learn the Signs dataset using the Tensorflow framework and achieved 84% accuracy. Implemented a ResNet (CNN) model using Keras and increased its accuracy to 86.8%.

Predicted daily rise/fall of Charles River Laboratories Inc. stock price with RNNs using LSTM.

Identified credit card frauds using Self Organizing Maps and used Artificial Neural Network to predict the probability of a customer being fraudulent.

Built a movie recommendation Boltzman Machine that predicts binary ratings and AutoEncoder that predicts numerical ratings (1-5) using Pytorch using Movielens dataset. Machine Learning

Built Random Forest and XGBoost models to predict safe/risky loans using LendingClub dataset.

Analyzed Social network ads. using models like Logistic Regression, Naïve Bayes, Kernel SVM, etc.

Performed advertisement CTR optimization using Thompson Sampling and Upper Confidence bound.

Created a bag of words NLP model to predict sentiment from restaurant reviews dataset. Optimization

Mathematically programmed several problems in scheduling, transportation and assignment, network flows, and supply chain optimization using Lingo, Matlab and Python. Hadoop and Spark Projects

Built a music recommendation engine using MLlib in Spark.

Analyzed Twitter feed stored in JSON for Presidential elections using SparkSQL. Database Design, Data Integration

Captured sales and purchase data of an enterprise from MS Sql Server, MySql, Oracle, PostgreSQL, and flat files to a target database using Talend and SSIS to improve analytical performance statistics.

Contact this candidate