EDUCATION

KAVANA VENKATESH

Boston, MA, ***** 857-***-**** adkaso@r.postjobfree.com

https://www.linkedin.com/in/kavana-venkatesh-56713a142/ https://github.com/KavanaVenkatesh Available: Jan - Aug 2021

Northeastern University, Boston, MA Sept. 2019 – Present Khoury College of Computer and Information Sciences Expected graduation: Dec 2021 Candidate for Master of Science in Data Science, GPA: 3.61/4.00 Related Courses: Deep Learning, Large Scale Parallel Data Processing, Supervised Machine Learning, Unsupervised Machine Learning, Data Management and Processing, Algorithms, Linear Algebra and Probability for Data Science. SJCE, Mysuru, India

Bachelor of Engineering in Electrical and Electronics Engineering, GPA: 9.27/10.0 June 2018 Related Courses: Advanced Calculus, Data Structures and Algorithms, Python and MATLAB, Signal Processing. TECHNICAL SKILLS

Languages/Databases: Python, R, C, SQL

Frameworks: Tensorflow, PyTorch, Keras, Scikit-Learn, Numpy, Pandas, NLTK Machine Learning: Supervised (Regression, KNN, RF, SVM, Decision Trees Naïve Bayes), Unsupervised (PCA, K-Means), XGBoost. Deep Learning: DNN, CNN, RNN, ResNets, LSTM, Boltzmann Machines, Autoencoders, Computer Vision Data Visualization: Matplotlib, Seaborn, ggplot2, Tableau, MS Excel, Shiny Application, Plotly Big data Processing: MapReduce, Apache Spark, Scala, Hadoop, AWS (EC2, EMR, S3), PySpark, Snowflake, PostgreSQL, Airflow, Koalas. PROFESSIONAL EXPERIENCE

Fidelity Investments, Boston, MA. Jan 2021- Present Data Science Co-Op Python, Deep Learning, Machine Learning, MySQL, PySpark, Cloud computing, AWS, Snowflake, SparkML, PostgreSQL

Interning in the core Data Science research and development team to build scalable machine learning and deep learning models for finance.

Extracted and transformed historical stock data with millions of observations using SQL and simulated expected Equity returns with the help of large-scale distributed collaborative filtering for missing data generation.

Built a risk model that predicts the risk associated with thousands of market factors by analyzing correlation and similarity using SparkML.

Building Variational Autoencoders and LSTM Models to simulate various portfolio returns using big historical data and pyspark. Indian Institute of Technology, Hyderabad, India May- Aug 2017 Research Fellow and Analyst Neural Networks, Machine Learning, Python, R, Data Processing.

Developed an optimized design of a Hybrid power generator using Artificial Neural Networks (Back propagation).

Implemented dashboards via Tableau and R Shiny app to visualize model’s results, to keep track of the circuit, and optimized model’s design using insights obtained from results. The design increased the output root mean square voltage by 29%, the best result so far. ACADEMIC PROJECTS

Northeastern University, Boston Sept 2019 -May 2020 Autonomous object detection using YOLO Python, Tensorflow, Computer Vision, Convolutional Neural Networks.

Developed an autonomous multiclass (80) object detection model using YOLO, after non-max suppression using Tensorflow and a Convolutional Neural Network pipeline.

Intelligent Movie Recommendation using Hybrid Deep Learning Python, PyTorch, Computer Vision, CNN.

Built a User-Item based Collaborative Filtering Recommendation System on MovieLens Data Set using a Hybrid Deep Learning Model with Restricted Boltzmann Machines and Autoencoders.

The model predicted the missing ratings of movies with a test loss of only 0.24. Customer Segmentation using Unsupervised Learning Python, Sklearn, Clustering, PCA, SOM, KNN

Created segments of customers based on their spending pattern to enable the Distributor to efficiently structure their delivery service and developed a model that allows to anticipate the purchases that will be made by a new customer.

Suggested optimal locations to establish and/ or expand the Business to maximize the profit. Twitter Based Sentiment Analysis of US Presidential Election RStudio, Twitter API, NLP, NLTK

Scraped unstructured data from Twitter using Twitter API, processed it and performed sentiment Analysis to extract insights about major issues on which candidates were basing their campaign and how it might impact their standing in the elections.

Obtained an overall accuracy of 78.4% in predicting the State-wise results.

Built an RShiny application to organize and present the results so obtained. PUBLICATION

Fault Analysis and Predictive maintenance of 3-phase induction motors using Machine Learning. Aug 2018 Best paper award at ICEECCOT- 2018 and the paper is published in IEEE Xplore digital library (https://ieeexplore.ieee.org/document/9001543)

Contact this candidate