Nithyasree Vivekanandham Pushpalatha

Gaze2vec: Learning embeddings from camera-based eye tracking data 1. Utilized OpenFace toolkit to extract raw gaze features 2. Classified emotions using SVM, Random Forest and CNN for AMIGOS dataset and compared with classifiers like SAXVSM and BOSSVS 3. Obtained latent embeddings from the pretrained Gaze360 dataset using TS-TCC (Time-Series representation learning framework via Temporal and Contextual Contrasting) 4. Analysed and visualised the learned embeddings using t-sne

Code sharing Tools and Processes 1. Restructured, packaged, and released software components as Conan packages 2. Migrated existing projects to a Conan-based dependency management workflow, simplifying the build and deployment process and reducing the risk of dependency conflicts 3. Worked with build systems like Make and CMake to integrate packages into existing projects and workflows 4. Set up automated continuous integration tests for newly released software modules using CI/CD

1. Predicted personality traits using the multi-modal dataset (videos and bio-signals like EEG; ECG and GSR) 2. Performed feature analysis using Fisher and RFE to identify key features that were strongly correlated with specific personality traits 3. Implemented binary classifiers for personality recognition in the individual setting using machine learning models like SVM, Random Forest and XGB 4. Encoded the training data into a high-dimensional space using Hyper-dimensional Computing (HDC)

1. Designed and implemented a Coolant Oil Monitoring System that utilized economical pH and Refractive Index (RI) sensors 2. Integrated the sensors into an IoT framework and designing a communication protocol using I2C to access the sensor data 3. Developed Python scripts to collect and process the sensor data and built a ThingWorx dashboard to visualize the data in real-time