Sign in

Engineer Data

Little Elm, Texas, United States
February 26, 2018

Contact this candidate


Amay Umradia

Little Elm, Texas ***** m: 469-***-****


Data Scientist / Machine Learning Engineer

Highly analytical and results-focused professional with trained in data science and machine learning techniques to gain deep insight into a broad range of business questions. Perform predictive analysis using historical data. Visualize data and write reports to effectively convey information to diverse audiences. Communicate with stakeholders to identify analytics requirements and select appropriate methodology to meet business objectives.

Possess six-years of hands-on experience in the telecom industry working with cutting-edge LTE features. Test algorithms using various parameters, observe trends, and adjust parameters using different values based on logical decisions to improve performance.

Core Competencies:

Stakeholder Communication

Prediction & Machine Learning

Statistical Analysis & Inference

Driving Solutions and Innovation

Team Leadership

Team Collaboration

Data Visualization Products

Data Gathering/Cleaning

Schedule Management

Technical Proficiencies

Machine Learning /Big Data

Sci-kit learn, Python, R,Pandas, Numpy, matplotlib, KNN Classifier, Logistic Regression, Linear/Quadratic Discriminant Analysis, Decision Trees, Random Forest, Support Vector Machine, PySpark, Hadoop, Tensor Flow, Deep Neural Network,PostgresSQL,pgadmin, Jupyter notebooks

Professional & Project Experience

Harvard University Udacity Nanodegree Program

Independent Projects

Completed two intensive certification programs involving extensive hands-on practicum work in data science, Supervised and Unsupervised machine learning, deep learning, and neural networks.

Key Projects:

Airbnb Pricing Prediction For New York House Listings (Fall 2016):

Performed Data Imputation using normal distribution and Visualization using matplotlib, seaborn to show seasonal variation of house pricing

Inferred hyper-parameters for Linear Regression with Ridge and Lasso technique to generate pricing predictions.

Enhanced accuracy of existing model by incorporating seasonal / holiday rate increases.

Airline Delay and Cancellation Visualization

Performed Data Visualization using matplotlib, seaborn to convey delay distribution and ranking the airlines based on the number of take-off flights

Various plots like flights path on maps, pie-charts comprised of different airlines, normalized histogram of delayed distribution and ranking them using priors, horizontal bar charts classifying on-time, short and long delays, variations of the delays with respect to the origin airport and for every airline.

Kaggle Competition – Automated Medical Diagnosis Classification (Fall 2016):

Classified imbalanced binary classes with over and under sampling techniques using Logistic Regression by tuning class weights, Linear and Quadratic Discriminant, Random Forest, and Support Vector Machines, Gradient Boosting Classifier to evaluate the best performance.

Data Science (Projects Utilizing Different Machine Learning Techniques), CS109A (Fall 2016):

Built risk assessment model to predicts financial risk for lender based on loan application information from the data set of Bank of America.

Observed land cover changes using satellite images to detect vegetation using various classification methods.

Used Natural Language Processing for sentiment analysis of Ford automobiles reviews.

Evaluated Linear Discriminant Analysis and Logistic Regression for changing income populations Boston. (time-series-analysis)

Identified disease subtypes based on biometric readings using KNN and Linear Regression.

Generate Visual Art using Tensor Flow (Spring 2017):

Successfully generated new image using style transfer technique of a style image into content image leveraging Deep Convolutional Neural Network and Tensor Flow on Amazon Web Service (Power GPU).

Big Data Analytics Projects using Various Big Data Tools, CS-63 (Spring 2017):

Implemented machine learning techniques using Amazon Web Services instances and FloydHub.

Expanded knowledge of Hadoop, PySpark, Amazon Web service, Microsoft Azure with H2O for Machine Learning, and Microsoft Azure for IoT.

Ericsson, Various Locations in the US

Solution Architect (Data Science Consultation)

Key Achievements:

Performed Data Exploration for Washing Machine IoT sensors using IBM bluemix IoT Platform and CloudantNoSQL Database to monitor anomalies in voltage and temperature using Spark and MapReduce

Utilized Amazon Web Service 8X GPU and Multiple-Linear Regression with Ridge, Lasso Regularization modelling technique to estimate download speed with feature engineering on various features to mitigate the multi-collinearity and biases in the result. This helps the customer to identify the problems and invest in necessary areas to improve network performance.

Using Data Visualization with Matplotlib and Seaborn analyzed predictive response of Dynamic Cell Capacity Algorithm based on commercial users’ traffic and resource utilization, taking peak and off hours into account

Using Logistic Regression, support vector machine (SVM) and Linear Discriminant Analysis (LDA) reliably classified the network into Low and High Overload case in order to take special measures for High overload sites to avoid business impacts of losing connected user.

Apply data mining and analysis on daily basis for various network performance metrics such as Accessibility, Retainability, Handover success rate and integrity using SAP application Business Object and Ericsson’s tool ITK.

New Product Introduction Engineer, 4/2011 – Present (RF Engineer Work Experience)

IoT (Internet of Things) trialed in AT&T labs. Presented feature demonstrations to client from initial concept through delivery using lab and field trials.

Key Achievements:

Expertise in translating highly technical algorithms and working with both non-technical and technical professionals.

Served as trusted partner to customers and promptly resolved issues to assist them in overcoming challenges.

Led new markets and stood out as one of only leads for ground-breaking carrier aggregation feature testing.

Won add-on sales and innovation idea awards

Education and Credentials

Master of Science in Electrical Engineering, 2010, University of Texas, Arlington, TX

Master Research Thesis, 2010 – Performance Evaluation of WiMAX, University of Texas, Arlington, TX

Data Science and Machine Learning Certification, Harvard University, Cambridge, MA

Deep Learning and Neural Network Nanodegree, 2017, Udacity

Contact this candidate