Data Scientist

Sai Kanaka Harsha Praneeth Nooli

Cell: 508-***-**** ac5ad6@r.postjobfree.com Worcester, MA, 01609 linkedin.com/in/praneethnooli Professional Summary:

I’m an aspiring Data Scientist currently pursuing MS in Data Science from Worcester Polytechnic Institute (WPI), MA. I have a total of 2 years of experience in Data Science with specific skills in Machine Learning, Python, R, SQL and Spark. I possess a keen interest in predictive analytics, and constantly work to produce interesting and useful insights. EDUCATION:

Worcester Polytechnic Institute (WPI), Worcester, MA Master of Science in Data Science, GPA 3.79/4.0 Aug 2016 - May 2018 Manipal Institute of Technology (MIT), Manipal, India Bachelor of Engineering in Electronics and Communication, GPA 8.85/10.00 July 2012 - May 2016 SKILLS:

Programming Skills: Python, R, C, MATLAB, Java

Machine Learning & Statistical Methods: Classification, Regression, Clustering, Feature Engineering, Bagging, Gradient Boosting, PCA, Bayesian Methods, Hypothesis Testing, Time Series Analysis, NLP, Feed Forward Neural Networks, CNN, RNN, LSTM Tools & Libraries: Jupyter Notebook, RStudio, SPSS, scikit-learn, numpy, pandas, scipy, keras, Spark ML, Spark Mllib Visualization Tools: Tableau, Matplotlib, ggplot2, Plotly Database & Big Data Technologies: SQL, MongoDB, Spark EXPERIENCE:

Data Science Intern, YAXA (Cyber-Security startup), Concord, MA May 2017 - Present

Built the login event modeling as a part of the company’s cybersecurity application in Python by using unsupervised machine learning algorithms like one-class SVM in order to detect a malicious login event.

Worked on behavioral modeling of users in Java and Python using real-time data in order to detect intrusions using models like Gaussian mixture models, one-class SVM, isolation forest etc.

Implemented LIME (Local Interpretable Model-Agnostic Explanation) to analyze feature importance for anomalous instances to find important features.

Performed suspicious connections analysis on Netflow data using Latent Dirichlet Allocation (LDA) in Spark for finding most suspicious events

Developed a LSTM based RNN in Keras to predict online browser behavior and detect anomalies. Research Assistant, UMass Medical School and WPI, MA Jan 2018 – Present

Using machine learning to classify EEG data based on different sessions, activities, and users. Data Science Intern, National Remote Sensing Centre, Hyderabad, India May - Oct 2015

Used Data Mining Techniques to process the rendered image from a satellite and detected the urban heat islands using CNN.

Created generalized SQL scripts for frequently used data operations to run on existing database to increase the performance of data retrieval.

Data Science Intern, Defense Research and Development Organization, Hyderabad, India Apr - Sep 2014

Developed an Intrusion Detection System using KDD dataset in Python to detect compromised packets based on various supervised machine learning classification techniques like Gradient Boosted Trees and SVM.

Published a research paper at RTEICT (Intl Conference on Recent Trends in Electronics, Information & Communication Technology). ACADEMIC PROJECTS:

My Rating: Handling Biased Reviews by Using Text Mining, WPI Aug - Oct 2017

Built a user-specific ML model in Python that identifies a user’s text review vs rating style and predicts the rating for this particular user given other user reviews as an input.

Used text mining and NLP to preprocess the data, applied bag-of-words, TF-IDF and Latent Dirichlet Allocation (LDA) to vectorize the cleaned textual data and used ML classification techniques to predict the ratings.

For users with less reviews, performed similarity clustering to identify which user-specific model applies to these users. Outlier Modeling of EEG Brain Signaling Data, WPI Oct - Dec 2016

Created a three-method framework in R to detect outliers from EEG Brain Signal data using Robust Principal Component Analysis.

Characterized noise and artifacts by using B-Spline and classified signal bands using Fourier Transform. Fake Review Detection, WPI Dec 2016

Built an online real-time detector in Python from the Yelp Dataset to predict whether a given review is deceptive or not.

Developed an anomaly detection model by using the elliptic envelope algorithm and a text classifier to detect fake reviews.

Contact this candidate