Urvi Patel

Phone: +1-226-***-**** Email: adw8oy@r.postjobfree.com

LinkedIn GitHub

Skills

Languages: SQL, Python, R Language, Octavo, JavaScript, NodeJS Web Technologies: HTML, CSS

Artificial Intelligence Tools: Regression, Classification, Clustering Models, TensorFlow, Keras, NumPy, Pandas, SciKit Learn, Natural Language Processing, Data Cleaning Packages: ggplot2, dplyr, Reshape2, plyr, pandas, numPy, TensorFlow, seaborn, sciPy, matplot lib, scikit-learn Soft Skills: Inquisitive, Teamwork and Collaboration, Organizational Skills, Time Management Experience

Dazzle Robotics Pt. Ltd., Gandhinagar, India June 2019 - September 2020 Worked as Senior Product Manager in an E-commerce company for electronics. Gained experience in interpreting and analysing data to improve sales of E-commerce website through digital marketing. Improved the sales rate by significant margin during my tenure.

Education

University of Waterloo, Waterloo, Ontario January 2021 - April 2022 Masters in Engineering with Artificial Intelligence and Machine Learning Specialization Under Electrical and Computer Engineering

Coursework: Data and Knowledge Modelling Analysis, Statistical Method for Data Analysis, Intelligent Systems Design, Quantitative Data Analysis for Management Science, Image Processing and Visual Communications, Project Management, Intelligent Sensors and Wireless Sensor Network Projects

o Real Time Chat App with NodeJS using Socket IO

− Runs on a local server developed using NodeJS.

− Technologies used for developing these projects are Nodemon, ExpressJS, Babel and Socket.IO.

− The front end will keep track of all the users’ sending messages.

− Messages sent from one client will be displayed in window of every client connected by common server. o Time Series Analysis of Stock using ARIMA in R

− Time series analysis using ARIMA model to predict stock prices for next three years on the historical stock data.

− Data pre-processing and analysing technologies employed are Tseries, Xts, Tidyverse, Forecast, Urca.

− Developed dashboard for live display of stock prices.

− The predicted stock prices had confidence interval of 99%. o Sentimental Analysis in Python

− Analysed and segregated IMDB reviews in positive and negative comments with NLP (Natural language Processing) Algorithms using TensorFlow, Keras, Sklearn.

− Implemented tokenization method for pre-processing of IMDB rating dataset to fed into Conv1D, bidirectional GRU and Conv1D with Regularization NLP models for prediction.

− The Conv1D network with regularization algorithm achieved accuracy of 89.3, which is highest in comparison with other implemented algorithms.

o Stock prices prediction using RNN in Python

− Employed RNN (Recurrent Neural Network) algorithms for future stock price prediction based on previous days prices 2015 to 2020 incorporating TensorFlow, Sklearn.

− Pre-processed data including variables volume, opening price and high and low prices were used to predicted stock price with LSTM, GRU and simple RNN algorithms.

− GRU proved to be more effective compared to other algorithms with MSE (Mean Square Error) 12.3, which mean predicted results were same as real values of those days. Certification (More details on LinkedIn)

o Structuring Machine Learning

o Improving Deep Learning Neural Network: Hyperparameter Tuning, Regularization and Optimization o Neural Network and Deep Learning

o Convolution Neural Network