Bryan Tamada

Prem Shah

Computer Science Graduate

About Me

I completed my Masters in Computer Science from Northeastern University, I am interested in the field of Software Engineering and Machine Learning. I completed my Bachelor's in Information Technology from Dharmsinh Desai University, India.
I have worked with ERT (eResearch Technology Inc.), a global company specializing in clinical services, as a Product Software Engineer Co-op. As part of the Syndication team, I was responsible for development of new modules for the LogPad and the SitePad app. I was also responsible for unit testing of the code base and resolving software defects assigned to the team. I also worked as Research Programmer and Research Assistant at Northeastern University where we develop Relational Agents which are computer agents designed to form long-term, social-emotional relationships with their users. Prior to starting my master's at Northeastern University, I was working as Software Engineer Intern at Institute for Plasma Research.

Work Experience

Research Assistant - Northeastern University (Jan 2018 - Dec 2018)

● Developed a robotic couple counselor using IrisTK framework and Furhat robot for promoting positive communication.
● Developed a program to do scientific analysis of speech in phonetics using c# and Praat to detect intensity raise and pitch drop.
● Used socket programming to pass head nod signals to Furhat robot for each intensity raise and pitch drop.

Product Software Engineer Co-op - e-Research Technologies(ERT) (Aug 2017 - Dec 2017)

● Contributed in building universal screen features and functionalities of an eCOA application using HTML5, CSS3, JavaScript.
● Developed tools using Apache Cordova for automating language validation and checking database capacity.
● Developed tools for enhancing performance, resolving defects and using Jasmine (JavaScript testing framework) for testing.

Research Programmer - Northeastern University (May 2017 - Aug 2017)

● Developed a relational agent that delivers a health & wellness intervention to African American men using C# and JavaScript.
● Evaluated health assessments to infer which health issues users are at risk for, and suggested potential solutions to those problems.
● Decreased system load and rendering time by 7% by refactoring code and redesigning MySQL schema.

Software Engineer Intern - Institute for Plasma Research(IPR) (Dec 2015 - Apr 2016)

● Designed a website that parses log files of software and stores it using MySQL to perform analysis of usage of software.
● Effectively designed web pages using PHP and HTML to minimize page-flow to reach a destination page and decrease bounce rate.
● Visualized analysis reports using Google charts and saved 15% revenue by recommending licensed toolboxes which needed renewal.

Projects

Brick Breaker Game & Pikachu’s Demise

Yelp Dataset

- Achieved 80% accuracy in developing a Naive Bayes Sentiment Prediction model using Apache Spark MLlib in Scala.
- Used many parallel data processing algorithms like Reduced-Side join and Secondary sort and ran whole project on AWS m4.large instance with 10-15 clusters.

Brick Breaker Game & Pikachu’s Demise

Category Classification

- Achieved 94% accuracy in predicting category of e-commerce products(Text Classification) based on their brand and description by developing a 2D convolutional neural network(CNN) model in Keras and also using Word2Vec embedding.
- Also tried using 1D CNN model with LSTM on top of it, also tried using Glove embedding.

Find out more

Brick Breaker Game & Pikachu’s Demise

Brick Breaker Game & Pikachu’s Demise

- Developed a mockup of the famous ‘Outbreak’ game and a platformer game using C++14, object oriented design and SDL library.
- Implemented a new gameplay mechanism and the ability to integrate difficulty settings using an editable config file.
- Built the executable file using Visual C++.

Find out more about Brick breaker game
Find out more about Pikachu’s Demise game

Financial Statements Analysis

Financial Statements Analysis

- Web scraped financial data of companies from SEC.gov website using python, requests library and lxml library.
- Designed a Machine Learning model having 85% accuracy on predicting companies that have durable competitive advantage by extracting data from annual reports and inferring income statement, balance sheet and cash flow statement for all 1500 companies.

Find out more

Diabetic Retinopathy Detection

Diabetic Retinopathy Detection(Kaggle Competition)

- Achieved 90% accuracy in diabetic detection by fine tuning VGG16 model in Keras & using stochastic gradient descent optimizer.

Link to Kaggle kernel

Pokemon

Analysis of Pokemon

- Analyzed pokemons based on it’s types, Max CP, Max HP and predicted pokemon based on it’s type combination and total power using Pandas library also visualized correlation between types and Mean HP and Mean CP using Matplotlib library.

Find out more

Quora Questions Pair

Quora Questions Pairs (Kaggle Competition)

- Designed a model having 60% accuracy from Quora Dataset to classify whether question pairs are duplicates or not.
- Visualized the most common words by building a data corpus using t-distributed stochastic neighbor embedding(t-SNE).
- Developed a model using Random Forest classifier, Xgboost boosting, Word2Vec, tf-idf and fuzzy strings techniques.

Link to Kaggle kernel

SpeakApp

SpeakApp

- Built a web app with Prof. Annunziato that connects students with native speakers via video chat (using MVC framework).
- Leveraged AngularJS for creating a front-end, single-page responsive application and Node.JS for RESTful web services.
- Migrated the database to MongoDB with Mongoose and added authentication using PassportJS.

Sentiment Analysis

Sentiment Analysis

- Achieved 55% accuracy in doing sentiment analysis using deep neural network.

Find out more