Sharanya Ganesh Prasad

- Data acquisition from drones, dataset generation by handling imbalance, augmentation and data cleaning. Dataset preparation: Image cropping, handling imbalance, augmentation. - Image data analysis and computer vision techniques using CNNs, transfer learning for image classification and object detection tasks. Experiments: Transfer Learning using ResNet, VGG, Inception, MobileNet, EfficientNet, Xception. Set-up: Training and inference in remote server via SSH and in AWS SageMaker

- Feature engineering and comparative analysis of various machine learning practices for anomaly detection in production using sensor data. - Evaluated already existing and state-of-the-art methods for feature extraction from the time-series data. - Two ways of feature reduction techniques experimented viz. Principal Component Analysis (PCA) and correlation metrics with the labels. - Documentation of empirical analysis results and provided the best strategy for classification depending on the scenario.

- Pioneered in communicating with cross-functional teams to understand the requirements and work on-site with the production team to develop a pipeline for quality detection in manufacturing using unsupervised learning. - Mathematically reformulated and modeled the regression problem as a classification problem. - Experimental evaluation of various algorithms for performing classification using neural networks and decision trees.

- Underwent technical training on ABAP and BODS. - Development of business applications in SAP environment for report generation and smart form creation.