Hello, I'm Arjun (AJ)

Developing a full-stack travel itinerary web application, enabling users to create, share, and discover day trip plans with social features. Designed and implemented normalized database schema with 8+ tables, built RESTful API with 20+ endpoints featuring JWT authentication, visibility controls, and with AI chat integration in progress.

Both an intelligent travel planning platform that will plan to use AI to recommend personalized travel itineraries, optimize routes, and provide real-time travel insights and a social platform where users can share and find popular itineraries from other users. A companion mobile app is also planned. Currently in active development.

Designed and developed a full-stack, serverless web application that converts text input into high-quality audio files using AWS cloud services. The platform includes secure user authentication, real-time text-to-speech conversion, and word frequency analytics, built with a serverless architecture that scales automatically with user demand.

Architected a comprehensive serverless solution leveraging 5+ AWS services, eliminating the need for traditional server management and infrastructure overhead. Designed and implemented a RESTful API architecture using API Gateway with two dedicated endpoints for text conversion and analytics queries, enabling efficient and scalable data processing.

Configured S3 for static website hosting to deploy the frontend application, ensuring fast and reliable content delivery. Implemented S3 bucket storage for audio file management with lifecycle management policies to optimize costs and automate file retention strategies.

Built a full-stack finance platform with portfolio tracking and AI-powered financial advisory capabilities. Implemented a retrieval augmented generation (RAG) system retrieving user portfolio data from PostgreSQL to augment Gemini API prompts for personalized investment advice.

A website that uses AI to analyze financial data and make predictions about the stock market. You create your own portfolio(s) and can view live stock data. The AI will analyze the data and make suggestions about based on your favorited portfolio. Configured to happen every day at 5 am EST. Additionally, users can chat with a chatbot to ask an AI financial questions with portfolio context obtained via user's favorited portfolio

Multi-Layer Perceptron (MLP) / Feed-Forward Neural Network for predicting heart disease risk using patient data analysis. Performed data cleaning, feature encoding, normalization, and train/test splitting to prepare structured medical data for supervised learning. Evaluated model performance on unseen data using classification metrics to assess predictive accuracy and generalization for healthcare risk assessment.

Uses a dataset from Kaggle to train the model, includes data preprocessing, where the data is cleaned and transformed. Model training, where the data is split into training and testing sets. And evaluation, where the model is evaluated on the testing set. A full breakdown is available through the public github repository linked below.