Tarini Sunkara

S.T.E.M. Educator at RoboThink Operations, specializing in innovative robotics projects that enhance student engagement. Demonstrated success in improving academic performance in mathematics and English as an Academic Tutor at Educally. Proficient in Python, with a strong focus on data analysis and machine learning techniques.

AI/Analytics Capstone Project | 02/2025 to 05/2025

University of Technology Sydney

• Developed a deep learning model utilizing Long Short-Term Memory (LSTM) networks to predict individual mental health indicators, such as depression levels, from longitudinal behavioral data, achieving an accuracy of 85%
• Pre-processed and conducted model training on the GlobeM dataset using the Adam optimizer and MSE using Python and TensorFlow for optimal performance.
• Assessed accuracy through a base model of logistic regression, which provided an accuracy of 66%, and was then leveraged as a base to develop the LSTM model.
• Collaborated within a team of six, utilizing GitHub for version control, and employing Agile methodology to oversee deliverables and milestones.

Machine Learning | 08/2024 to 11/2024

University of Technology Sydney

• Constructed and evaluated neural network models using the UCI Wine Quality dataset for binary classification, achieving an AUC of 87%.
• Improved model performance of multi-class classification through hyperparameter tuning with Keras Tuner, accomplishing a validation accuracy of 63% from 54%.
• Leveraged Python libraries such as Scikit-learn and pandas to apply data preprocessing techniques, such as normalization, label encoding, and one-hot encoding, to increase input quality.