When I started programming when I was 9, I was building for myself. I built 200+ websites and apps by the time I was 15, when I entered the industry and started working fulltime. I still love to build projects in my free time.
This web app fetches news articles from hundreds of sources (such as Bloomberg and the Washington Post) and allows them to be searchable and sortable.
It's built in Next.js and React, and uses popular React packages like react-select. It's also completely mobile responsive.
I created all of the styling for the site, and no UI frameworks were used.
It takes advantage of Next.js' server-side capabilities by fetching articles on the server and caching scraped data.
This project was built with Handlebars and SASS, with source code compiled and minified to a static site with Webpack.
I built the visualization functionality first in Rust and PHP to generate images on the server-side, but later opted for a solution in HTML5 Canvas.
I built the Cube class to store 3D cube data, process moves, and manipulate the faces, and I built the solver program from scratch as well.
The web app is mobile responsive and is built as a PWA using WorkBox. As a PWA, the app works on mobile devices, even when offline.
Includes latest case count numbers, news on the virus from more than 100 countries, and latest media briefings from the WHO.
I built and deployed this site in under a week. It reached the front page of Hacker News and has had thousands of users.
Python cronjobs are used to update data on the site through a PHP server.
The design and UI was written in completely custom CSS without any framework(s).
Jupiter is a completely machine-learning based AI that achieves the 2048 tile (wins the game) almost 100% of the time. I built and published Jupiter in September 2020, and wrote three articles about the project on my blog.
I used web workers, multi-threaded JS, and webpack minification to increase performance, speed, and load-times.
I implemented machine learning techniques including the MCTS (Monte-Carlo Tree Search) algorithm in JavaScript.
I also wrote a technical article about the AI algorithm used in Jupiter, which reached the front page of Hacker News and was read by thousands of developers.
I identified a common transcription error in archived NY Times articles, which appeared over 300 times in just three notable articles, including the first article in the Pentagon Papers release. My program fixes the errors automatically with ease.
I wrote the program in JavaScript to be run on both the web and on a server with Node. The Google Ngram Dictionary and commonality rankings were used to help fix errors in words.
I built and deployed a site with a demo of the program using Vercel with GitHub actions.
I am currently liaising with a NY Times archivist to see how this program can be of help to their team.
I built this site for the Standard (standard.asl.org), showcasing a timeline of articles, infographics, and more.
The site was viewed by thousands of students, parents, and faculty in a matter of hours and received praise from the principal and the head of school.
The site got second place in National Scholastic Press Association Best of Show Award (Spring 2020, for News/Feature Package).
I personally received merit from the National Scholastic Press Association for my design and programming work on this site.
Daily Developer Jokes posts developer humor ever day on the website and on the DEV community.
The Daily Developer Jokes bot, which posts the daily jokes, is the first ever bot on the DEV community, and uses the DEV API.
The bot uses a PHP server to generate joke banners and store images generated with Puppeteer and Node.js.
To post jokes daily, a Python cronjobs system was implemented to interface with the DEV API and send email updates to maintainers.
The site uses GitHub pages, and jokes are stored as markdown and compiled to a static site using Jekyll and Ruby.
I built HTML5 Rocket as a set of responsive site templates for use without attribution.
I designed all of the sites without any CSS or UI frameworks.
All of the templates are easily downloadable and customizable from the HTML5 Rocket Website.
