Shubhang Sharma - Portfolio

Education

Georgia Institute of Technology, Atlanta, GA - BS, Computer Science, Graduated Honors May. 2023

07/2023 - Ongoing

05/2022 – 08/2022

Implemented a custom changelog service to increase observability of NCR's codebase by building and deploying the service using Go, Google Cloud, and Github Actions.
This increased transparency and improved communication with clients about version updates and new features in the codebase.
Developed a Github bot to improve the quality of commit and pull request messages. In addition, developed a Slackbot to send updates to NCR's Slack workspace.
Constructed a web page using React to allow users to filter, search, and view all changelogs.
Developed a GraphQL API in Golang to query and filter change log messages from a Google Cloud Platform Big Query table and deployed the server using Kubernetes.

08/2021 – 08/2022

Developed production-ready, testable, and scalable NLP models.
Utilized PyTesseract on a high-performance computing cluster to automate the conversion of over 1.8M pages in scanned PDF files to excel data, resulting in annual cost savings of $200K.

01/2020 – 08/2020

Led end-to-end development and deployment of a Flask web app on a Ubuntu Linode Server.
Developed an NLP model to process over 1M Yahoo Finance conversations, predict future prices, market sentiment, and generate a historical analysis for stocks.
Created a summarization algorithm to gather and summarize law and technology news for consultants at EY in a daily feed. The algorithm reduced the corpus size by 85% and organized the news by industry sector. All of these features were integrated into the Flask app and deployed on a website.

08/2020 – 08/2021

Developed ML-based systems leveraging clinical data to identify patients with Liver Lesions and other anomalies.
Designed optimization algorithms to solve class imbalance problems with sparse data to create effective datasets for machine learning modeling, resulting in 4x effectiveness over previous manual process.
Built and managed ML-engineering teams, leading complex cross-functional projects, and mentoring freshmen & sophomore students to uplevel technical skills and domain knowledge.
Sourced, onboarded, and managed a remote team of 2 medical annotation specialists, labeling 360K clinical records via custom annotation tool to train deep learning models.
Oversaw development of feature extraction and modeling system for large CT-Scans & DICOM datasets, allowing for the rapid training and deployment of new models for business partners.

05/2020 – 08/2020

Worked on deploying an LSTM inventory management model at Avantor to handle both overstocking and understocking and predict cycle times, reducing overstock issues for high movement SKUs by 80%.

03/2020 – 12/2020

Developed an ML-based approach leveraging medical data to identify biomarkers for Alzheimer’s Disease.
Research Published in ACS Nano, March 25, 2022. Link to Publication