Projects

Social journal made to keep in touch with my friends across the globe. Check out the Site!

Organic traction and paying users.

AI phone agent for restaurants that I went full-time on in December 2025. Check out the Site and Demo Video!

Spent four months building it, meeting 50+ restaurant owners, and initiating pilots with restaurants and jewellery stores.
Stopped after realising voice AI models weren't where I wanted them to be yet — which they probably are now, as of August 2026.

Built the alumni tracker site for TAMU Sling Health. Check out the Demo Video and Sling Health TAMU site!

I designed and developed the entire front end.

Won Best Use of Notion with a campus app that connects toots (tutors) and tootees (tutees) while making study room booking smarter and more social. Check out the Demo Video and Devpost!

Won 3rd out of 700+ with an AI-powered interactive DJ app that personalizes music curation. Check out the Demo Video and Devpost!

Used computer vision, generative AI conversation topics, and real-time music APIs.
Built the frontend using React and Next.js, enabling a dynamic and responsive user experience.

Built an incubation game studio on Fortnite achieving 500+ CCU and 11,000+ plays.

Won 1st out of 800+ with a full-stack interview prep app that analyzes speech, expression, and eye contact. Check out the Demo Video and Devpost!

Full stack application to help prepare people for interviews by analyzing speech, emotions, and eye contact, while asking potential interview questions about the role that they are applying for.
Backend using Pytorch, OpenCV, and SpeechRecognition, built, trained, and fine-tuned 4 ML models for facial recognition, expressions, eye tracking, and speech-to-text.
Integrated with OpenAI's API for data processing, question generation, and improvement summarisation.
Displayed an animated Unity frontend model through React.

Built a website with two different visual styles through CSS manipulation without JavaScript libraries. Check out the Site!

The site you're on right now - I revamped it in 2026 to be faster, cleaner, and a lot nicer to look at.

The thinking behind it (newest first):

Why rebuild it... again?
Honestly, AI changed the math. The kind of refactor that used to eat a whole weekend now takes an afternoon, so I finally had the budget to redo the foundation properly. The old site was a Create React App single-page app - one big JS bundle that rendered everything in the browser. I migrated it to Next.js with a fully static export, so every page is pre-rendered to plain HTML (better for load speed and SEO) and still hosted on GitHub Pages with no server to babysit. I also pulled all the content - projects, experience, links - out into typed data files, so adding a project like this one is now literally just editing an array.
Mobile-first, accessibility-first nav
On phones the menu opens into a full-screen overlay with a proper focus trap, so keyboard and screen-reader folks don't get stranded behind it. The desktop dropdowns got the same once-over: smoother open animations and no weird seams. The kind of stuff nobody notices unless it's broken, which is exactly why I sweated it.
Ideas that survived the rebuild
Plenty of the old version's ideas were good enough to keep - just reskinned. Here's what I wrote about them last time:
Design Process + Auto Open Tab
When you landed here, you probably noticed this tab popped open right away. That's on purpose! I picked up some great insights from a talk by a Google UX Designer, who highlighted the importance of showing not just what a website offers, but also the thinking behind its features. Initially, I thought about creating a separate page to share these thoughts, but then it struck me - why not make it immediately visible? Easier for you to find and keeps the site looking neat!
Interactive Navigation Buttons
Turns out, not everyone is a fan of traditional navigation bars (who knew, right?). So, I spiced things up with some fun, interactive buttons that you'll find on our homepage and about page. They're designed to make your experience not just smoother but more satisfying.
Light Mode Toggle
I started off with a dark theme because, honestly, I just love how it looks. But, hey, it's not for everyone! So, I added a light mode toggle. Now, you can switch it up anytime you like, making the site more accessible and comfortable for all eyes.
Accordion
I've also put in an accordion layout on the projects and experience pages. It's a sleek way to pack in information without overwhelming you. Just click and expand the section you're interested in!

Design inspiration sources!

Placed 4th out of 50+ with a TensorFlow/Keras CNN drawing classifier built on Quick, Draw! data and custom classes. Check out the Devpost!

Built an ML model from 200,000 data points to classify passengers as leisure or business and recommend margin improvements. Check out the GitHub!

Led multidisciplinary teams across industry-sponsored pitch competitions to solve real-world problem statements for AI, robotics, aviation, and security concepts.
These competitions allow me to work with a diverse team to solve real-world problems and practice presenting to industry professionals. I have led each team and worked on concepts and design, as well as the video filming and editing.

The competitions (newest first):

Sandia National Labs: AI/ML — Sep 12, 2023
Designed "ScanX," deep-learning software that identifies illicit contraband in x-ray content and extrapolates individual items into an easy-to-understand grid. Check out the Presentation!
Sandia National Labs: AI/ML — 3rd Place — Apr 16, 2023
Designed "VTOLtally Awesome," a long-endurance VTOL-capable aircraft that could be used for search and rescue as well as reconnaissance.
Los Alamos National Laboratory: Nuclear Security — Oct 23, 2022
Designed "Spy Kids," a spy-bot swarm that detects nuclear radiation and other potential threats. Check out the Presentation!
National Security Agency: Informational Viz — 4th Place — Sep 25, 2022
Designed "Mediator," an extension and app that evaluates the credibility of social media posts through machine learning. Check out the Presentation!