Shreyas Patel

Montréal, Québec 365-***-**** adz1w5@r.postjobfree.com shreyaskpatel

Key Skills

Programming Languages: Java, Python, C++, C, JavaScript. Database: MySQL, PostgreSQL, MongoDB, Neo4j. Web Technologies: HTML5, CSS3, jQuery, Flask, Angular, PHP. ML Libraries: TensorFlow, Keras, PyTorch, scikit-learn. Developer Tools: GitHub, Git, Jira, Jenkins, VS Code, IntelliJ, Eclipse, Anaconda, Google Colab, AWS. Concepts & Methodologies: Agile Software Development (Scrum), Object Oriented Programming, Software Development Lifecycle

(SDLC), Functional Programming, Unit Testing, Continuous Integration and Continuous Deployment, Machine Learning, Transfer Learning, Data Augmentation, Natural Language Processing (NLP). Education

Concordia University (GPA 4.01/4.30) Jan 2022 – Present Master of Applied Computer Science. Montreal, Québec Sardar Vallabhbhai National Institute of Technology Jul 2016 – Jul 2020 Bachelor of Technology in Computer Engineering Surat, India Experience

Samsung R&D Jan 2021 – Nov 2021

Software Engineer Delhi, India

• Developed and maintained Samsung Smart TV app registration and verification portals utilizing a microservices-based architecture, while implementing robust continuous integration and continuous delivery practices.

• Worked on supervised learning models to forecast whether apps would pass or fail testing based on app metadata, resulting in a reduction of over 30% in resource expenditure for the QA team.

• Achieved Professional Level certification in the Samsung Global Software Competency Test, achieving a top 3% ranking among peers by showcasing exceptional problem-solving, algorithmic expertise, and programming skills. Sardar Vallabhbhai National Institute of Technology Aug 2020 - Dec 2020 Research Intern Surat, India

• Conducted a research project called "Emotion Analysis," which involved classifying tweets based on 13 emotions.

• Employed various sampling techniques to address imbalance and integrated pre-trained word embeddings using GloVe.

• Constructed a BiLSTM Neural Network Model for tweet classification, resulting in an impressive 172% improvement over the baseline model. My research was subsequently published by IGI Global. Doctor AI May 2019 - Jul 2019

Machine Learning Intern Bengaluru, India

• Detected fraudulent credit card transactions, employed preprocessing techniques to handle missing data and eliminate noise from the data, addressed the dataset's imbalance, and built a Random Forest Classifier that boosted performance by 105%.

• Estimated steel production rates, used dimensionality reduction techniques to address the overfitting problem mitigate by high-dimensional data and applied the XGBoost model, resulting in a notable 150% enhancement in accuracy. Publications

“Impact of Balancing Techniques for Imbalanced Class Distribution on Twitter Data for Emotion Analysis: A Case Study,” IGI Global, June 2021.

Projects

Face Mask Detector Python, Data Augmentation, PyTorch, scikit-learn. Jun 2022

• Implemented a Convolutional Neural Network (CNN) system to analyze facial images, accurately detecting mask usage, assessing correct wearing, and classifying mask types with 88% accuracy, improving public health safety protocols.

• Used K-fold cross-validation, image augmentation techniques, and optimized the CNN architecture for image classification. The system successfully reduced bias related to age groups by 40%. FreelanceIOT Advanced Java, API Integration, Play Framework, Mockito, Actor-Based Programming Apr 2022

• Designed a web application that enables real-time analysis of freelance website data through APIs. The application utilizes advanced programming concepts like functional programming, lambdas, streams, and asynchronous programming with futures, Websockets, and Google Juice for dependency injection, resulting in 35% improvement in execution time. Sentiment Analysis of Twitter Data Python, API Utilization, Tensorflow, Keras, Flask, Data Analysis May 2020

• Implemented Twitter API to extract location-based tweets, incorporating emotions and categories; data stored in Neo4j database for analysis and insights, addressing class imbalances and utilizing diverse text preprocessing techniques.

• Utilized GloVe word embeddings to train Bidirectional LSTM and SVM models for sentiment analysis and category classification, providing accurate insights into user mental states and community behavior.