•Given an image of a steel sheet predict the type of defect from one of the four types of defects defined.
•pre-processed the Images using gamma correction, contrast enhancement and added Run-length encoding.
•Built a UNET model for localizing and classifying the surface defects on the steel sheet.
•Achieved a Dice coefficient of 0.82 for classifying images.

• Predicted the scene of a given image using a neural network.
• Prepared, augmented using ImageDataGenerator in Keras in the Intel scene classification dataset for classification.
• Built a Deep learning Network to Classify images into 6 categories with an accuracy of 93 percent.

•Predicted the rating that a user would give to a movie that he has not yet rated and Minimize the difference between predicted and actual rating.
•Merged movies with users and their rating in a single data frame, Computed User-User Similarity matrix and Movie-Movie similarity matrix, Featurized the data by selecting the top 5 ratings given by similar users to a particular user and also top 5 ratings given to similar movies with respect to a particular movie.
•worked with different machine learning models and Achieved a Root Mean Square Error of less than 1.08.

•Predicted the number of pickups of a given Location Using a Regression Model.
•Cleaned the data based on several features, got different regions based on clusters, created the different baseline models for different months, taken the previous five values as the feature, included the exponential average forecasting feature and Fourier Transform Features.
•applied the different models like linear Regression, Random Forest and XgBoost Regression models and achieved a mean absolute percentage error of less than 12%.

• Implemented End to End Deep learning Network to predict the degree of the steering angle with the help of 25 minutes of Footage.
• Used convolutional neural networks (CNNs) to map the raw pixels from a front-facing camera to the steering commands for a self-driving car.

•Predicted the tags based on the content that was there in the question posted on Stack overflow.
•Predicted as many tags as possible with high precision and recall.
•Performed NLP using Bag of Words, Tf-idf Vectorizer and Classified the Tags Using Logistic, linear regression models.
•Achieved Hamming Loss of less than 0.0004 and a macro-f1 score of less than 0.09.

• Identify which questions asked on Quora are duplicates of questions that have already been asked.
• Achieved a Log-loss of 0.36

• Performed sentiment analysis to classify the polarity of the review given by the Amazon user , given the textual reviews and related features of the product.
• Performed Exploratory Data Analysis, data cleaning, data visualization and text featurization (BOW, tfidf, word2vec).Buildseveral ML models like KNN, Logistic regression,SVM,Random Forest,GBDT,LSYM(RNNs)etc.