DevanshBeyond the code.

Overview

Tripzy brings the fragmented parts of trip planning into one focused workspace. It pairs generative AI with live travel data so a destination, a budget, and a schedule turn into an itinerary that actually makes sense.

What it covers

• Generative itineraries tailored to budget, pace, and interests
• Flight fare prediction layered into trip planning
• Weather-aware recommendations and booking links in one flow
• Full-stack architecture with a PostgreSQL data layer

Overview

Rio was built to understand the moving parts behind a social product: identity, media, feeds, and the small interactions that make a network feel alive. It is a complete authentication and content pipeline from database to interface.

What it covers

• JWT-based authentication and protected application routes
• Post creation with Cloudinary-backed media uploads
• Likes, comments, and social interaction flows
• REST API designed around clean Express controllers

Overview

MiniRTV is a research-driven build that explores how computer vision models can understand body pose, preserve garment structure, and produce a believable try-on result without specialist hardware.

What it covers

• Pose and body landmark estimation for garment alignment
• Deep-learning-based image synthesis experiments
• A research notebook documenting model behaviour and limits
• A practical testbed for computer-vision workflows

Research build · code available on GitHub

Overview

What started as a simple question — who is actually on top? — became an exercise in data cleaning, aggregation, and visual storytelling. The project turns a noisy public dataset into a clear, comparative view of streaming behaviour.

What it covers

• Scraped and normalised global streaming data
• Aggregation with Pandas across artists and markets
• Visual comparisons of the most-streamed artists
• A foundation for larger big-data experiments

Research build · code available on GitHub