Upper Extremity Evaluation Assistant
A web-based clinical decision-support prototype for occupational and physical therapists performing upper-extremity evaluations.
Tech Stack:
-
Codex for system architecture and coding
-
GitHub for source control
-
HTML, CSS, and JavaScript for the frontend
-
Vercel for public deployment
-
Vercel serverless function for /api/evidence
-
Bright Data SERP API as an optional evidence-acquisition layer
Transformer Model Efficiency
Fine-Tuning Transformers for Toxicity Detection: A Sentiment-Informed Approach.
This NLP project highlights the effectiveness of sentiment-informed training in enhancing toxicity detection while providing a scalable solution for real-world applications. By incorporating sentiment, the model gains additional context to better understand the intent behind an author's online post or comment.
Reddit Finance Posts
This project explores the intersection of behavioral science and social media analytics, specifically utilizing data scraped from Reddit. Built NLP models to predict what makes Reddit posts more engaging using scraped social media data. Compared BERT, Bag-of-Words, transformer-based approaches, and baseline regression to evaluate how language and behavioral signals influence engagement.
Co-founded a women’s health platform designed to create a safer space for open health discussions and access to expert-informed content. Built the foundation for an AI-powered content moderation system, later using the platform as the real-world use case for my research into toxicity detection with machine learning and transformer models.
Perceive
Perceive is an assistive shopping app that uses real-time grocery item recognition to guide users through stores with precision. Unlike conventional tools that flood users with unnecessary information, Perceive AI focuses solely on items the user actually needs: those on their shopping list.
Through clear directional audio guidance and a screen-reader-friendly interface, the app ensures a smooth, frustration-free experience. With built-in voice command functionality, users can interact hands-free and with full autonomy.
Research Paper RAG
Absorbing and skimming through hundreds of thousands of research papers is an impossible task for a single person, even with current LLM tools. During a one-day Hackathon, our team developed an agent that enables users to summarize over 100,000 research papers in their field and extract key information through prompts. This was achieved by building a Retrieval-Augmented Generation (RAG) system leveraging open source LLMs.
