Projects

Osler

3rd Place at RuHealthHack - AI-powered mobile application for hospitalist onboarding efficiency. Built with Flask backend and RAG system using Google Gemini for intelligent information retrieval. Features include real-time hospital guidelines access, staff scheduling integration, supply location mapping, and administrative tools for notifications and KPI analysis. Implemented vector embeddings for semantic search and automated content summarization.

Python Flask RAG Vector DB Google Gemini AWS

Aura

Enterprise AI analytics platform for retail management with conversational intelligence. Features multi-step business investigation capabilities through Snowflake data warehouse integration, real-time analytics dashboard, and natural language query processing. Implemented advanced SQL optimization, automated report generation, and predictive analytics for inventory and sales forecasting.

Python Snowflake LLM SQL Analytics

Mesh

Enterprise knowledge management platform that transforms unstructured data (documents, meeting transcripts, whiteboards) into queryable knowledge graphs using AI-powered ingestion pipelines. Implemented OAuth2 authentication, real-time collaboration features, and automated content summarization with vector embeddings for semantic search capabilities. Won first at hackathon hosted by Rutgers X Fiserv.

React OAuth2 Vector DB AI Real-time

Vault

Secure team collaboration platform built with Flask backend and encrypted file storage. Features include AES-256 encrypted messaging, secure file upload/download with integrity verification, task management system, and role-based access control. Implemented secure authentication, session management, and audit logging for enterprise compliance.

Python Flask AES-256 RBAC Security

Contextualized-Multi-Modal-RAG-Feedback-Loop

Advanced RAG system with contextual learning capabilities using GPT-3.5 and GPT-4. Implemented multi-modal data processing, personalized query optimization through machine learning feedback loops, and dynamic context window management. Features include semantic similarity matching, query expansion, and automated relevance scoring for improved search accuracy.

Python GPT-4 RAG ML NLP

BATMAN: Protect your City

Civic intelligence platform providing real-time crime and safety data for urban residents. Built with Node.js backend and MongoDB for data persistence, featuring interactive crime mapping, incident reporting system, and automated SMS notifications via Vonage API. Implemented geospatial data processing, real-time alerts, and community reporting features for enhanced public safety awareness.

Node.js MongoDB Geospatial SMS API Real-time

GIST

Intelligent GitHub repository analysis tool using web scraping and OpenAI integration. Features automated code analysis, dependency mapping, documentation generation, and project insights extraction. Implemented GitHub API integration, natural language processing for code understanding, and automated report generation for repository assessment and onboarding.

Python GitHub API OpenAI Web Scraping NLP

PLANA

AI-powered business plan analysis platform with multi-model machine learning capabilities. Features comprehensive financial modeling, market analysis, risk assessment, and strategic recommendations. Implemented natural language processing for document analysis, automated SWOT analysis generation, and predictive modeling for business outcome forecasting.

Python ML NLP Financial Modeling Analytics

GPT 3.5 4 Devs

Open-source GPT enhancement framework integrating multiple Python libraries and APIs. Features web scraping automation, document processing, file system interaction, and Streamlit-based user interface. Implemented modular architecture for extensible AI workflows, automated data pipeline creation, and interactive dashboard for AI-powered task automation.

Python OpenAI Web Scraping Streamlit APIs

Helios

AI-powered haptic navigation device integrating computer vision and natural language processing for object detection and spatial guidance. Developed embedded system using Python, OpenCV for real-time object recognition, NLP for voice command processing, and haptic feedback algorithms. Features proximity sensors and machine learning models for mixed-object detection within close-range environments.

Python OpenCV NLP Haptics Embedded

ARGlasses

Augmented reality e-commerce application for virtual eyewear try-on experience. Built with Swift and ARKit for iOS, featuring real-time face tracking, 3D model rendering, and virtual product placement. Implemented computer vision algorithms for facial landmark detection, photorealistic rendering pipeline, and interactive UI for seamless virtual shopping experience.

Swift ARKit Computer Vision 3D Rendering iOS

Cone

A PHP application that allows users to share all their social media, contact information, and external links by QRCode. It has a login/signup page for users to create an account, an automatic QRcode generator, and a sharable link.

Web PHP

Inexpensive Enhanced One Factor Authentication

Advanced logistic regression model to predict the likelihood of a user login being malicious based on various parameters.

Python Machine Learning Security

ELO Algorithm Based NFL Predictor

An ELO Ranking system to rank NFL teams using historical data and then predicting output of games.

Python Sports Analytics

Congress Connect

Congress Connect is an app that allows people to learn about congressmen's views on various subjects, based on their tweets using sentiment analysis and machine learning.

Swift Sentiment Analysis UI/UX

Around-The-Pipe Audio Based Water Tracker for Home Residents

This research presents a revolutionary approach that utilizes audio to measure the amount of water going through a household pipe at any given time with supervised machine learning models.

NLP AI Python Research

CourseTurtle

A startup aimed at making an easier way to analyze student and teachers school course schedules. It allows for users to create their own accounts and share the data in an extremely accessible way.

Web PHP

QuickEstimate

An app that can scan various household items and classify them with Machine Learning, to then produce a pricing estimate for each of those items. The purpose of this app is to make it easy for insurance estimators to create a pricing estimation. Won 3rd Place in 2021 NJ-07 Congressional App Challenge. I had the opportunity to present the project to representative Tom Malinowski.

Swift CV UI/UX Python

ryanrana._