Souradip Roy
Phone: +1-609-***-****; Email:adiub5@r.postjobfree.com
**** **** *** *, ********** Village 117, Fargo, ND 58102 Permanent Address – 0 Michaels Drive, Chelmsford, MA 01863 1
Career Objective
Build a long-term career as a Deep Learning and Machine Learning Researcher and keep flourishing in the field. Additionally, I would like to enhance my knowledge and professional skills including but not limited to machine learning in a stable and dynamic workplace. I am always keen to impart the skills that I had acquired over the years for practical application. Education
PhD in Computer Science: Aug 2019 – Present, Computer Science, North Dakota State University, Fargo, USA. MS in Computer Science: Aug 2017 – July 2019, Computer Science, North Dakota State University, Fargo, USA
B. Tech in Computer Science: Aug 2011 – Jul 2015, Computer Science, West Bengal University of Technology, Kolkata, India
Technical Skills
Languages: Python
Frameworks: Keras, Pandas, Scikit-Learn, Scipy
Databases: MySQL, MSSQL, POSTgre-SQL
Others: Node, React, Django, HTML5, CSS3, Bootstrap IDE: Jupyter Notebook, Visual Studio Code
Research Experience
Graduate Student (PhD), North Dakota State University, (Aug 2019-present)
• Hands on experience in NSL-KDD dataset handling using Pandas.
• Multiclass and Binary classifier, like Feed Forward Neural Network, Recurrent Neural Network and Statistical models, creation for intrusion detection in IoT networks using Scikit-learn, Python and Keras.
• Actively involved in the concept and experiment design of the classifier.
• Experienced in handling features with collinearity by grouping the independent variables based on variance influential factor values of each feature.
• Feature Selection using a combination of XG-Boost feature scoring and Variation Inflation Factor.
• Extensive experience for dimensionality reduction with the help of PCA and Stacked Autoencoder.
• Performed sampling using Random Under Sampling with Replacement, Tomek Links and SMOTE
(knn) oversampling to create a balanced dataset.
• Experienced in working with ensemble learning like stacking and boosting.
• Knowledge about various evaluation methods like Confusion Matrix, Precision-Recall curve, Receiver Operating Curve, Accuracy, Area Under Curve. Graduate Student (Masters), North Dakota State University, (Jan 2018 – May 2019)
• Performed binary classification of Wisconsin Breast Cancer dataset. Souradip Roy
Phone: +1-609-***-****; Email:adiub5@r.postjobfree.com
1085 17th Ave N, University Village 117, Fargo, ND 58102 Permanent Address – 0 Michaels Drive, Chelmsford, MA 01863 2
• Extensive knowledge in creating classifier models like Support Vector Machine, Logistic Regression, Naïve Bayes, Decision Tree, Dense Neural Network.
• Used knn oversampling technique to create a balanced dataset for the minority class.
• Compared all the classifiers based on values of Confusion Matrix, Accuracy, Area Under Curve to conclude the best performing model.
• Experience with Chest X-ray image dataset.
• Extensively used Keras to build a Densenet121 model.
• Performed transfer learning to train the model from imagenet to classify 14 types of diseases.
• The evaluation of the model has been carried out based on Confusion Matrix and Accuracy.
• Experience in handling of gene expression dataset having feature set in thousands.
• Performed feature extraction and classification using Stacked Autoencoder classifier and Dense Feed Forward Neural network.
Co-op Experience
Jul 2019 – Present: IT Intern at Upper Great Plains Transportation Institute
• Development of Stored Procedures, Functions in MSSQL.
• Web Services using Express and Nodejs backend.
• Frontend development using ReactJs.
• Using Geolocations for different features in the application. Work Experience
Jun 2016-Jul 2017: Application Developer and Team Leader at Rplanx Technology Pvt. Ltd
• Developed Dynamic Web Page Applications using Django Web framework.
• Web Services for Android Applications using Django RESTful framework.
• Directly consulted with the clients regarding their project requirements and features they need.
• Google Search Result Web Crawler using Python.
• Deploying Projects to EC2 instance using Elastic beanstalk.
• Created frontend pages using Bootstrap, HTML5, CSS3. Publications
Roy S., Li J., Choi B.J., Bai Y., A Supervised Anomaly-Based Intrusion Detection Mechanism for IoT Networks. Manuscript in preparation.
Roy S., Li J., Fog-Cloud Transfer Learning Intrusion Detection Model for IoT Networks. Manuscript in preparation.