No Risk, No Story! 

Worked in a group of 6 students to develop a web application that can track and predict the likelihood of Software project failure based on its features during its development. Undertook feature engineering using the NumPy Python library, and developed a logistic regression model using the scikit-learn Python library to predict the outcome of a software project. Used Flask API to integrate the ML model with the Node.js and PostgreSQL server. Final project achieved the 2nd highest grade in the cohort.

Explored various types of Continuous Learning methods applied to Vision Transformers(ViT). Applied Adaptive-Distillation-of-Adapters(ADA) to HuggingFace pre-trained ViT for image classification. The approach improved the ViT prediction accuracy by 20% on 20 tasks trained sequentially.