• Transformed the variables by using data manipulation techniques like One-Hot Encoding and performed an Exploratory Data Analysis to see the impact of variables over Sales.

• Used ML models such as Linear-Regression, Lasso, XGRegressor. verified model performance along with techniques such as PCA, GridSearchCV for hyper-parameter tuning.

• Time-Series analysis/prediction with traditional ML model ARIMA and also LSTM.

• Built Deep Learning CNN (Convolutional Neural Network) model to classify Cats and Dogs.

• Validated data augmentation techniques to overcome the challenge of fewer data images (20 images of each category) and there was an improvement in the model performance

Implemented Object detection model using the jetson-interface package, deployed on NVIDIA Jetson-Nano board.

• Face recognition model training and deployment on the Jetson-Nano/raspberry-pi 4

• Regression Problem, prediction of target variable in seconds time that the car needs to pass the test

• Performed EDA, label Encoding, PCA for dimensionality reduction

• Developed XGBoostRegressor model.

• Classification Problem, Identification of the level of income qualification needed for the families in Latin America
• Able to achieve accuracy of 94.27% with RandomForestClassifier

• Developed NLP pipeline for text casing, cleaning, tokenizing using NLTK.
• Used GridSearchCV and KFold for optimal parameters of LogisticRegression.

Implemented object detection model using the jetson-interface package, deployed on NVIDIA Jetson-Nano board Face recognition model training and deployment on the Jetson-Nano/raspberry-pi