A quick overview of some projects I’ve worked on:
2024 – Continentle
This was a super fun app to make. It’s a geography puzzle game where you guess a country or state (I call it the “continentle”) each day, within a limited number of tries. There is a practice mode as well, where you can play infinitely. After guessing, you can compare your guess to other players with a leaderboard, powered by MongoDB. I was mostly inspired from fantastic apps like Wordle and Travle
- Tech Stack: Next.js, MongoDB, react-globe.gl, shadcn, Tailwind CSS, deployed via Vercel.
- URL → continentle.com
2024 – CloudStream
I worked with a team to create a robust video streaming storage system, designed for efficient uploading, processing, and streaming of videos.
The goal of the project was to implement a distributed object storage system (similar to Amazon S3). The system allowed for highly available, and efficient distributed storage and retrieval of objects. To achieve this, we built custom load balancers, storage nodes, and data routers, integrating Zookeeper for distributed coordination and fault tolerance.
- Tech Stack: Java, Docker
2024 – Petite DB
This was an incredibly comprehensive Database system created from the ground up. Some aspects included buffer management, file handling, undo functionality, logging, metadata, query parsing, record management, and transaction processing. In the end, a SQL query (in natural language) could be parsed, optimized, and executed within the system, interacting with the underlying storage and transaction mechanisms.
The hardest part about this project was rollback functionality, ensuring correct undo and redo operations.
- Tech Stack: Java
2023 – songseekr
This app lets you get in-depth musical information about your Spotify library, allowing you practice a musical instrument along with your own songs. It also serves as a UI for your library, as well as letting you search any song you can think of. If you don’t want to connect your account to Spotify, or if you don’t use Spotify, you can also just search in guest mode.
This one still needs a bit more polish, but was one of my favorites because it totally added value in my own life. I use this app constantly, allowing me to practice guitar along with songs that I actually like.
- Tech Stack: React.js, Node.js, Material UI, OAuth 2.0, deployed with Vercel.
- URL → songseekr.com
2023 – Fault-Tolerant Distributed System
This was a fault-tolerant system that used a Zookeeper-inspired algorithm to handle leader election, allowing for automatic consensus among the distributed nodes. Additionally, gossip-based UDP heartbeats were utilized to constantly detect failures.
- Tech Stack: Java
2023 – Regression Line
This project won the first prize in our university hackathon! Essentially, it was a tool that provided data visualizations of stocks. The project showed important metrics including unusual moves taken from 10-Q financial statements for the individual stock ticker. Instead of combing through 100’s of SEC documents, users could instantly see outliers and trends in SEC filings.
- Tech Stack: Python (for data analysis) and React.js (for data visualization).
2021 – Smart City
I collaborated on an intricate AWS application packaged in Docker. It was essentially a cloud based site that allowed users to report/find issues in a city. Most time was spent making the comprehensive AWS system, along with CloudFormation, Lamba Functions, and an API Gateway. Finally, we also allowed the system to be packaged in Docker, allowing for easy portability and setup.
Personally, I found this application super valuable as it was an introduction into building comprehensive system infrastructure. It was fascinating to plan an entire cloud-based, scalable system and make each piece work together.
- Tech Stack: Java Spring Boot, CloudFormation, API Gateway, Lambda, RDS, S3, Secrets Manager, Docker, JavaScript/CSS.
- Github → Smart City
2022 – MediMind
My 2022 hackathon project was an app that helped patients remember to take their medication. It would notify patients at scheduled times and track their adherence to prescribed medications. It had an effective login system and simple UI for scheduling reminders.
- Tech Stack: Python (Flask) backend and React.js frontend.
2021 – Document Store
This project was a Java-based search engine optimized for efficiently handling document storage both in memory and on disk. There was also built-in undo functionality (based on a LRU eviction strategy), JSON-based storage, and robust performance for large data sets.
- Tech Stack: Java (Custom data structures (B-tree, Trie, Heap, Hash Table))