Research Projects

LLM for Medical Action Ontology Curation

Innovating a language model algorithm using OntoGPT, PubTator3 and LangChain for efficient medical literature analysis, enhancing rare disease curation for the Medical Action Ontology (MaXO) database that contributes to treatments and research on rare diseases.

Under supervision of Prof. Peter Robinson

Optimizing Word2Vec Algorithm

Authored and developed a Python package `wn2vec`, optimizing Word2Vec, NLP techniques, tested using TensorFlow and Genism ML algorithms and spearheaded the manuscript as the first author. [https://github.com/TheJacksonLaboratory/wn2vec]

The Robinson Lab

NASA Research

• Lead a research team in parsing and interpreting extensive NASA solar observatory datasets, aiming to demystify, and forecast solar anomalies.

• Developed ML models such as CNNs and LSTMs using TensorFlow and PyTorch, analyzing over 1 million solar images (6.5+ TB) to enhance the predictive modeling of solar events like coronal holes.

SSD/HHD Research

• Building Machine Learning Models to predict SSD & HDD failures using Alibaba Inc servers data under the supervision of Dr. Chandranil Chakraborttii at Trinity College in partnership with Goodwill Computing Lab from North Eastern University

• Utilized Python to analyze SSD and HDD data from Alibaba, Inc. servers and built machine learning models to predict SSD and HDD failures.