PORTFOLIO
Featured Current Research:
MARVIS - The Universal Browser Agent API
The Vision: AI For Anything and Everything
Marvis embodies the concept of "API for Everything" by transforming any web-based service, application, or interface into a programmable API endpoint. Instead of being limited to services that provide official APIs, Marvis can interact with any website or web application as if it were a native API.
What’s the value?
Marvis eliminates the fundamental limitation of the modern web: the API gap. While the internet contains millions of useful services, most lack programmatic access. Marvis bridges this gap by making every web service programmable, turning the entire web into one massive, interconnected ecosystem. This democratizes access to web functionality and enables developers to build integrations that were previously impossible or prohibitively complex.
Learn more here!
Want to just jump in?
Visit my GitHub.
AI Healthcare Projects
-
Personalized Contraceptive Recommendation Bot
National Conference of Undergraduate Research awarded LLM-powered GenAI system that provides personalized contraceptive recommendations based on a women’s health data.
-
Diabetes Diagnose using Machine Learning
Predicting diabetes risk based on inputted user health metrics, demonstrating how AI can make preventative healthcare more accessible.
-
DNA Classification Algorithm
Explored the fascinating field of bioinformatics by developing a classification system for DNA sequences, transforms genetic data into numerical formats suitable for analysis.
-
Pneumonia Detection using Chest X-ray Images
Built an automated diagnostic tool that identifies pneumonia in chest X-ray images through deep learning addressing a critical healthcare need
-
Neural Interfaces for Healthcare Applications
Implemented a neural signal processing system using EMOTIV BCI technology that translates brain activity into game controls, exploring its applications to healthcare.
-
LSTM Networks for Automated Cardiac Arrhythmia Detection
Technical analysis of the use of recurrent neural networks (RNNs), specifically long short-term memory (LSTM), to enable temporal modeling for improved detection
AI Integration Projects
-
Automated Livestock Quantification System
Implemented a YOLOv5-based object detection system for automated livestock quantification from aerial imagery. Engineered custom data augmentation pipeline to enhance model robustness against variable lighting and occlusion scenarios.
-
Ticketmaster Event Management System
Real-time queue optimization system leveraging predictive analytics and dynamic resource allocation algorithms to streamline event entry processes. Implements crowd density monitoring via computer vision for adaptive capacity.
-
Political Bias Detector
KNN-powered algorithmic system that can automatically classify political bias in news articles. This tool helps readers understand potential slants in their news consumption and promotes media literacy.
-
Spotify Personalized Music Assistant
AI system utilizing natural language processing and collaborative filtering algorithms to generate contextual music recommendations based on user preferences and acoustic feature analysis.
-
Smart Meal Prepping Platform
App that helps make the transition to independent living smoother and more enjoyable through organized, efficient meal management. Using ingredient recognition and preference learning, it suggests personalized recipes that match your pantry and tastes.
-
Google Smart Home Mirror
Ambient computing interface integrating Google Gemini multimodal AI with computer vision and voice recognition. Features contextual awareness through environmental sensors and personalized interaction through user recognition algorithms.