Associate VP

• Oct 2023 - Present

Data Scientist Intern

• Sep 2021 - Feb 2022

• Localized events in videos utilizing the PyScene Detect API and extracted 20 frames with 2560 features for captioning. • Processed training data by performing caption tokenization and padding to establish a 15-words vocabulary list. • Improved caption generation time by 13% using greedy search with an Efficient Net B7 and Seq2Seq model, featuring a BiLSTM encoder and LSTM decoder. • Boosted BLEU and METEOR by 4.4% and 15.1% higher than baseline, achieved scores of 0.78 and 0.34, respectively Show less

Machine Learning Intern

• Sep 2020 - Jan 2021

• Processed Sloan dataset by resizing and Otsu binarizing images with a threshold of 190 pixels to generate binary images. • Executed performance assessment using MSE (49.20) and PSNR (31.85), demonstrating the accuracy of the methodology. • Constructed a robust pipeline by integrating VGG16 and K-means clustering for galaxy classification, achieving a silhouette score of 0.82. • Processed Sloan dataset by resizing and Otsu binarizing images with a threshold of 190 pixels to generate binary images. • Executed performance assessment using MSE (49.20) and PSNR (31.85), demonstrating the accuracy of the methodology. • Constructed a robust pipeline by integrating VGG16 and K-means clustering for galaxy classification, achieving a silhouette score of 0.82.