Worked for 3.9 years as AWS and Backup administrator at Infosys Ltd.
Graduate Projects: Title: Geolocation Clustering Objective: Implement k-means in SPARK and use it for geo-location clustering on various datasets of spatial locations. Observations: Implementing k-mean algorithm with Great circle method takes more time compared to Euclidean distance method and computation time increases with the number of clusters.
Title: Adult Census Income Objective: To predict the annual income range (<=50k or >50k) of US population, by analyzing various parameters of adult census income data and building a predictive model. Machine learning models used: Logistic regression, Naive Bayes, K-nearest Neighbors Result: Built an annual income predictive model with accuracy score of 81% using Naive Bayes algorithm.
Title: Fully connected neural network Objective: To build a 3 layered fully connected neural network using basic python programming and analyze the performance of model with varied parameters for Iris and MNIST datasets. Result: Achieved model accuracy of 89% for optimal combination of parameters.