Developed a regression model by using random forest regression of h2o to predict number of bugs in software in production. Able to achieved around 0.83 r squared.

Technologies/Tools: h20, distributed random forest regression, steam, R.

Domain: Software

Developed aclassification model topredict the customers who can upgrade mobile.

Able to achieved 79?curacy.We have used GBM of h2o.

Technologies/Tools: h20, GBM, steam, python, pandas, numpy, SMOTE.

Domain: Telecom /Sales

Developed a classification model to predict the recurrence of hepatitis C disease after the completion of treatment. Able to achieved the accuracy of 91%.We have used ensemble learning by using random forest, GBM and SVM.

Technologies/Tools: R, SMOTE, ensemble learning, random forest, gbm, svm, baruta, pca, caret

Domain: Pharmaceuticals

Developed a classification model to reduce the false positiveinstances MBIalgorithm. Able to reduce the false positive cases by 7%.We have fine-tuned the already used random forest and by using various techniques of variable selection.

Technologies/Tools: R, baruta, leaps, pca, SMOTE, under sampling, logistic regression, random forest, svm, gbm

Domain: Pharmaceuticals

Developed an association rule mining model to find patterns in the variables - which accident conditions frequently occur together. We have used apriori algorithm of R.

Technologies/Tools: R, apriori

Domain: Insurance

Developed a sentiment analysis model for comments written by employee in yearly engagement survey. This is polarity based model to classify the comments in Highly Positive, Positive, Neutral, Negative and Highly Negative.

Technologies/Tools: Python, VADER

Domain: HR

Developed a regression model to predict the emission of various cars as per Environmental Protection Agency (EPA). Able to achieved the accuracy of 78%.We have used ridge regression of R.

Technologies/Tools: R, ridge regression

Domain: Automobiles

Developed classification model to predict whether person can be prospective Bank Customer. Able to achieved the accuracy of 79%.We have random forest of R and also used clustering to get better prediction.

Technologies/Tools: R, SMOTE, K-means, random forest

Domain: Banking

Individual Contributor

Developed a cold start accessories recommendation model for telecom organization Technologies/Tools: Python, Knn, pandas, numpy and matplotlib Domain: Telecom / Sales

Individual Contibutor

Developed a personalized accessories recommendation model for telecom organization Technologies/Tools: Python, Kmeans, pandas, numpy and matplotlib Organization: HCL Technologies Domain: Telecom / Sales

Individual Contributor

Working on text classification (natural language processing) by using transfer learning of google pre trained word2vec (GoogleNews-vectors-negative300) model. Technologies/Tools: Keras, CNN, LSTM, Python, pandas, numpy, word embedding, nlp. Domain: Telecom / Sales