Georgia Institute of Technology, Atlanta, GA
B.S. in Computer Science & Stamps President's Scholar (Expected Graduation: May 2027)
Hi, I'm Vidhi, a computer science major and Stamps President's Scholar at Georgia Tech with concentrations in artificial intelligence and systems & architecture. My prior research spans computer graphics, robotics, genomics, metabolomics, and machine learning, tackling challenges in spaceflight, health diagnostics, and real-time monitoring. With a background in computer graphics and embedded programming, I hope to design cutting-edge systems for real-time visualization and intelligent device integration.
Education
Work Experience
Bioinformatics Intern @ NASA Gene Lab
Led development of research proposal on RNTL & RORc genes' impact on lipid accumulation in the liver of astronauts via the disrupted circadian rhythm during spaceflight. Utilized Python and R to analyze Gene Lab data of mice euthanized on the ISS. Research accepted to ASGSR conference on Capitol Hill. Part of NASA AI/ML analysis working group.
Robotics Intern @ Preemadonna
Developed augmented reality features for the “Nailbot”, a smart nail art robot, using computer vision, real-time image processing, and robotic control features. Built the Preemadonna website by integrating Keyboard APIs, responsive SVGs, and server-side logic.
Machine Learning Researcher @ San Diego Supercomputer Center
Trained Random Forest machine learning model with blood metabolites for early diagnosis of feline mammary cancer. Studied pathways with close relationship with human breast cancer. Use of bioinformatics tools such as Weka, Stringdb, Metaboanalyst. Analyzed SMILES values for metabolites and PaDEL descriptors.
Bioengineering Researcher @ Boolean & Turakhia Lab
Conducted analysis of Sars-Cov-2 data to find sites in the genome that may be under selection using UCSC Genome Browser, Python, and Linux. Developed a machine learning model for arrhythmia detection based on ECG readings implemented on a TinyML device for real-time analysis using embedded programming
Projects
Biothings APIs
Collaborated on development of BioThings APIs which have over 24 million requests monthly. Utilized Google Analytics 4 on Biothings API to track user requests. Wrote Python parsers to classify drugs by unique pharmacological class in API using Python and JS. Optimized number of API calls made by Biothings API by processing qualifier information.
Nailbot
Developed artificial reality features of the Nailbot.
EyeReplace Chrome Extension
Developed EyeReplace, combining a Python backend with a JavaScript Chrome extension using the Beam Eyewear SDK for real-time eye-tracking and vocabulary simplification through gaze analysis and LLMs, with data stored in MongoDB. Implemented Google Authentication for role-based access, enabling educators to track student progress
Publications
Implementation of Machine Learning-Based System for Early Diagnosis of Feline Mammary Carcinomas through Blood Metabolite Profiling
Vidhi Kulkarni, Igor F Tsigelny, Valentia Kouznetsova
Feline mammary carcinoma (FMC) is a highly aggressive neoplasm with limited diagnostic reliability using conventional methods like imaging and fine-needle aspiration. This study identified key metabolic pathways, such as alanine, aspartate, and glutamate metabolism, along with critical oncogenic genes, including ERBB2, EGFR, and PDGFA, using MetaboAnalyst 5.0 and STRINGdb for metabolomic and proteomic analysis. A machine learning model trained on blood metabolite data demonstrated 85.11% accuracy, highlighting its potential for early, biomarker-driven FMC diagnosis. Read More