Developed a smart data pipeline with TensorFlow and Scikit-learn to ensure data homogeneity from heterogeneous data sources.
Developed a modular script to streamline data queries for multiple relational databases without SQL.
Collaborated with researchers in a transfer learning research using the Bidirectional Encoder Representations from Transformers (BERT) and Word2Vec models, to extend supervised learning and natural language understanding capabilities.
Collaborated with two research labs within the Digital Futures complex in the development of a perception algorithm using the spaCy Rule-Based pattern matching workflow, and the BERT model to detect Pro- or Anti-vaccine perception.
School of IT (Civic Tech Lab)
Conducted machine learning research to detect similarities between 911, 311, and social media data for easy crisis response with Cincinnati open data and Twitter academic API.
Developed the API module for an operational picture tool used to map local beliefs towards the COVID-19 pandemic with the Django framework and Natural Language Processing for opinion mining.
Integrated Named Entity Recognition and Word2Vec into situational awareness dataset from CrisisNLP for managing health crises.
Worked both independently and collaboratively in recording experimental findings in the weekly laboratory meetups to analyze and interpret reports from various works of literature.
