May 2021 – Jul 2021 · 2 mos · Bengaluru, Karnataka, India

• Classification of Digital and Handwritten Images
• Used Data Augmentation techniques to generate synthetic images of Digital and Handwritten text with a color background.
• Build a CNN model architecture consisting of Conv, MaxPool, and Dense layers supporting images of different widths.
• Used TF-Records for the training datasets and, checkpoints technique to avoid repeated training from scratch.
• The final model recognized over 1000 unseen images into digital and handwritten, giving a prediction accuracy of 90%.

Aug 2020 – Dec 2020 · 4 mos

May 2020 – Aug 2020 · 3 mos · Hamilton, Ontario, Canada

• ElmSr Web IDE for ElmJr: A Projectional Editor
• Created IDE supporting data types such as int, bool, string, let..in, definitions, etc. by the use of Zipper data structure.
• Replaced a data type with another type via a pop-up menu, new let..in & new definition by a single button tap.
• Edited variable names and updated calls simultaneously, supporting the Undo and Redo commands.
• Received Letter of Recommendation for the constructive contribution to the project.