AI Engineer

• Feb 2023 - Present

Deep Learning Engineer

• Nov 2020 - Feb 2023

Data Scientist

• Jul 2019 - Oct 2020

Member of Core ML team working on a Document Data Extraction product Sara Below are some of the highlighted work that skips many interesting contributions that are hard to explain in a sentence. * Developed a novel Rectangular Contour Detection algorithm for scanned Documents. Uses Hough Lines and performs better than opencv’s rectangular contour detection * Developed document templatization and data extraction algorithms that achieves 95+% data extraction accuracy on templatized… Show more Member of Core ML team working on a Document Data Extraction product Sara Below are some of the highlighted work that skips many interesting contributions that are hard to explain in a sentence. * Developed a novel Rectangular Contour Detection algorithm for scanned Documents. Uses Hough Lines and performs better than opencv’s rectangular contour detection * Developed document templatization and data extraction algorithms that achieves 95+% data extraction accuracy on templatized documents * Developed a Document Embedding model for clustering and classifying documents. Uses crawl embeddings and BiLSTM, and the documents are clustered based on embedding cosine-similarity. * Converted most of the codebase to OOD and statically typed and wrote API validators and decorator based encryption and logging modules. * Developed OOD document handling pipelines for images, pdf, docx, eml, msg, html, xls and csv formats where the data is converted into our standard format for processing while preserving all the positional info. Show less Member of Core ML team working on a Document Data Extraction product Sara Below are some of the highlighted work that skips many interesting contributions that are hard to explain in a sentence. * Developed a novel Rectangular Contour Detection algorithm for scanned Documents. Uses Hough Lines and performs better than opencv’s rectangular contour detection * Developed document templatization and data extraction algorithms that achieves 95+% data extraction accuracy on templatized… Show more Member of Core ML team working on a Document Data Extraction product Sara Below are some of the highlighted work that skips many interesting contributions that are hard to explain in a sentence. * Developed a novel Rectangular Contour Detection algorithm for scanned Documents. Uses Hough Lines and performs better than opencv’s rectangular contour detection * Developed document templatization and data extraction algorithms that achieves 95+% data extraction accuracy on templatized documents * Developed a Document Embedding model for clustering and classifying documents. Uses crawl embeddings and BiLSTM, and the documents are clustered based on embedding cosine-similarity. * Converted most of the codebase to OOD and statically typed and wrote API validators and decorator based encryption and logging modules. * Developed OOD document handling pipelines for images, pdf, docx, eml, msg, html, xls and csv formats where the data is converted into our standard format for processing while preserving all the positional info. Show less

Teaching Assistant

• Aug 2017 - Jun 2019

Electrical Design Engineer

• Sep 2017 - Dec 2017