Featured Projects
User Behaviour and Content Simulation
Developed a ML model to try to solve the problem of behavior simulation (Task-1) and content simulation (Task-2), thereby helping marketers to estimate user engagement on their social media content as well as create content that elicits the desired key performance indicators (KPI) from the audience.
Natural Disaster Image Detection and Segmentation
Developed a convolutional neural networks (CNNs) to identify and segment damaged infrastructure elements from satellite imagery with 95% accuracy, enabling faster disaster response and resource allocation by 60%
Medical vs. Non-Medical Image Classifier
A machine learning pipeline to automatically classify images as medical or non-medical from a mixed set of inputs(URLs, PDFs, etc). It is a a Python-based Computer Vision CLI application with web scraping capabilities.
Mars Crater detection
Developed and trained a convolutional neural network (CNN) to accurately detect and classify Martian surface craters from satellite imagery, achieving 96% precision and supporting automated planetary surface analysis.
17th Century Spanish Text Recognition
Build a model based on transformer architectures for optically recognizing the text of 17th Century Spanish data source. THe model is able to detect the main text in each page, while disregarding other embellishments.
CMS Event Classification
Modeled a Masked Transformer Autoencoders for event classifications at the Compact Muon Solenoid (CMS) a general-purpose detector at the Large Hadron Collider (LHC).