Ｒａｆｓａｎ

AI Agent Skills Marketplace

Full-stack marketplace where experts publish AI-agent skills and agents enroll. Integrated Stripe for payments and expert payouts, with skill search, filtering, and agent collaboration.

• Published clawvard — a public npm CLI and MCP server for autonomous skill discovery
• MCP server with browse, enroll, and list tools plus Bearer token auth
• Built with TypeScript, Next.js, and Supabase

Real-Time Competitive Trivia · AI-Generated Questions

Competitive trivia for iOS with 3v3 team mode and live gameplay. Questions powered by an LLM pipeline on Groq's inference API.

• Flutter frontend with Supabase backend
• Real-time multiplayer sync and team-based scoring
• Swift native components where needed

PocketCare is a bluetooth low energy (BLE) based solution which enables smartphones to send and receive anonymous beacon signals. It checks the distance between a smartphone and another beacon (or smartphone running PocketCare) to see if they are close to each other (less than 2m).

Competition: IBM Call For Code 2020 (Covid-19 Track)

My role: iOS Engineer

• Helped redesign the UI using Sketch (Fun fact: I designed the logo!)
• Implemented the iOS beacon library for close encounter detection
• Added background transmission/scanning using CoreBluetooth+Overflow
• Implement modified Hamming encoding/decoding algorithm in Swift.
• GeoFencing for identifying On/Off campus, CoreData(Storage), ARKit for calibration
• Developed plans for future releases, including implementation details on features.

Also made a quick little video for IBM submission
Checkout our work and our phenomenal team!

A platform dedicated to revolutionizing 21st-century education. Whether you use it to find tutors at a bargain price or ask questions and study "live" with others, Torble is built to raise a climate of possibility in education.

With features that allow dynamic collaboration and community Q&A, the platform is built to boost your academic prowess.

Founder + Built the first prototype

• Implemented Firebase cloud functions in JS to automate tutor-student matchmaking.
• Integrated Stripe(express and connect) API in swift
• Added video chat capability with Agora SDK.

PubSpek is a mobile app I built my sophomore year to address a problem a lot of people (including me) deal with - fear of public speaking aka glossophobia. It's one of the first application that gives you feedback on your presentation and helps you pinpoint core issues related to public speaking. I used Microsoft's cognitive API and IBM's face API to recognize facial expressions from videos, and then compare that with the appropriate one from the text. Based on that comparison the user would get a score with suggestions on how they can improve.
Challenges: I had to get the server involved in order to process videos in real time (extracting frames every second). It wasn't a perfect solution but the end result was better than expected. I was able to prove to myself that I am capable of exploring and learning different technologies as long as I stuck with it.

AR version of the popular mobile game Plague Inc!. I barely knew Apple's new ARKit framework when I entered the competition, but by the time we were done I realized the potential it has in the future.

Mobile application I built for Erie Hack competition during my freshman year. The app collects data from all the water suppliers in users region, and then shows user information regarding their suppliers. For suppliers it would provide with aggregated data from users, helping them to solve problems before they enter critical phase (like Flint water crisis)

An IoT toothbrush that uses self contained sensors and camera modules to detect a user's oral health status. Hardest part of this project was to get the camera module to work. I also didn't setup my PI beforehand, to make things easy I went through each room in our department to find a HDMI cable. Then figured out the SSH route using a handy tool called LanScan.

For my CSE442 (Software Engineering) course, we had to team up and come up with a product that we would be working on throughout duration of the course. The course was run exactly like a company, with meetings every week with our PM, using project management tools like ZenHub, and learning the agile development process. We developed this cool Piazza like forum with social media features.

Final project for my CSE341 class, we had to implement a famous clustering algorithm in MIPS. I looked everywhere to find a tool that would help me visualize the algorithm in real-time (back then my ML knowledge was very limited). I was finally able to figure it out and had a lot of fun coding it.

One of the best hackathons I've ever been to. The vibe was insane, but more than that I was fortunate to work with a team from two different countries, Australia and Canada.
We had a change of plan because we couldn't get the hardware necesary to build what we originally planned. Instead, since all of us had prior knowledge in iOS dev, we put together an app that turns your face into a QR code.