Urlefy is a free URL shortener for business and professional needs. 15K+ URLs are shortened and we have 160+ customers with 99% uptime.
 

Place your lengthy URLs and convert them into self-working short links. Urlefy helps you in URL shortening and analytics.
 

I Lead this project by creating a team of 5 people. 2 developers including me, one designer, and two marketers. I worked on brainstorming, product requirements, designing, developing, marketing, and maintaining the project.

It took us 2 months to create this version of Urlefy. The designed architecture is microservice-based and completely scalable, read about it here. Closely working with different clients and evolving with feedback loops and a supermassive managed clean code based. Created an in-house headless client management system for blogging. Designed heat maps and point maps to show exact geo-locations of users.

This project enhanced my leadership, communication, and teamwork. I learned about product development with the client-first approach. The passion to wake up each day knowing I'll be working on a challenging feature today.

Problems faced

1. Designing a well scalable and secure architecture. Emphasizing security.
2. The shortened links should be quick to load, SEO-friendly through all sharing platforms, and should collect all possible information about the users opening them (geo location, timestamps, mobile device, browser, and more).
3. As the links and data increased for each client, the dashboard loading time increased exponentially. Needed better algorithms and database management to reduce this many factors.
4. People kept misusing the site by shortening porn sites and illegal phishing sites. Google kept banning our redirection server.

How I overcame them

1. I studied microservice architecture, picked up best practices from tens of patterns, and then designed Urlefy.
2. Reduced redirection time to a few milliseconds by caching frequently used links and distributing the search in clusters.
3. Reduced load time of dashboard analytics by a factor of 80%. Did this by using better data structures, priority queuing next load requests, and background running tasks on the rabbit queue.
4. Fixed the site down issues within 6 hours by removing the phishing sites and figuring patterns and blocking such matched URLs.

Stack

Frontend: Nextjs, chartjs, mapbox, google OAuth, markdown for blogs

Backend: Nodejs, expressjs, mongoose, multer, passport, razorpay, mailgun, aws-sdk

DevOps: Cloudinary, AWS EC2, AWS Elastic IPs, Heroku, Mailing server

Blogging CMS: Strapi, Heroku, PostgresSQL

Database: PostgresSQL, MongoDB, AWS S3