About Me

Scuba Diver

I've been scuba diving for four years and logged over 160 dives. I am PADI divemaster, drysuit, EAN, and rescue diver certified. I've dived in more than ten countries and been on three liveaboards. I typically dive in the Monterey, CA area and help guide divers and assist with courses in the area.

RF Engineer

I became interested in RF engineering when I was first employed by a Department of Homeland Security lab working to build millimeter wave radar for airport security. I continued on to work in millimeter wave communications at a startup called Starry. . I worked at Apple. for two and a half years on cellular RF front end. Currently, I'm building RF modules for drones at startup Skydio.

Hacker

I attended my first hackathon as a high school senior in 2014 and was hooked from then on. In the last 6 years, I have attended fifteen hackathons, winning eight awards, including gold medals at BU and UPenn and bronze medals at Columbia and NYU. I am also the co-founder of Northeastern's Hackathon, HuskyHacks. My hackathon projects can be found in my Devpost portfolio. Outside of hackathons, I have also spent free time building UAVs, a plasma speaker, and a high frequency cryptocurrency arbitrage bot. I currently have a dive site mapping website in the works.

Check Out Some Projects I Made!

Click a Picture to See it in Action!

Pong Over Laser Link

At Columbia University's hardware hackathon, I worked with Anthony Bisulco, Max Connor, and Jay Shome to build a laser communication link. Pong Over Laser Link implements a prototype of LIFI using 1mW red Keyes laser modules to transmit at a frequency of up to 3.6 kHz. Utilizing phototransistors on the receiving end of the link, we are able to reproduce the transmitted signal. In order to maximize efficiency and reliability, we developed a custom packet format. After developing the data transmittion link and protocol, we implemented a game of Pong on LCD screens controlled by joysticks to demonstrate our data link. Check out our Devpost page or Read More.

CubeBot

Last summer, my friend Eric Liu and I designed, constructed, and programmed a robot to solve Rubik's Cubes. The project involved a lot of challenges on both the hardware and software sides. In order to come in under the size of almost all existing such robots, we custom modeled and 3D printed most of our components. On the software side, we implemented an existing cube solver on the raspberry pi, used OpenCV to capture the color of each square, and passed a set of instructions to an Arduino, which solves the cube using a time optimization algorithm. Our best solve time is 35s. Check out a video of it here or Read More.

Mood-Disco

At Cornell's fall hackathon, I teamed up with Anthony Bisulco, Max Connor, and Eric Liu to build a conversation sentiment analysis and feedback tool. Mood Disco listens a conversation using a Raspberry Pi microphone and converts the speech to text using IBM's Watson API. The words are then analyzed for eight different possible sentiments using a neural net we created and trained to recognize 60,000 words using a 300 word training set. Two moving averages of sentiment in the conversation are then created and audio and visual feedback is displayed using a speaker connected to Raspberry Pi and an RGB LED strip connected to the raspberry pi GPIO pins. Check out our Devpost page or Read More.

Smart Bike Rack

At PennApps XIII, I worked with Lyle Cheatham from Waterloo and Neil Chen from Columbia to create a Smart Bike Rack. In order to combat bike theft, we wanted to build the first ever Internet of Things enabled bike rack. We made good use of UPenn's rapid prototyping on site by creating our entire structure in Solidworks and building it using a laser cutter. In order to detect even slight tampering of the electrified bike like cable, we implemented a voltage divider feeding into a differentiating amplifier, which was read on a regular iteration by an analog Arduino input. An Arduino Leonardo operated a touch screen with which the user could interact, which then sent the user's phone number and PIN code to a Raspberry Pi via serial. In the event of tampering, the Pi texts the user and calls police via Twilio, and lights and buzzers are set of on an Arduino Mini communicated with via I2C (Inter IC). Check it outhere.

Ottr

Check out this travel website we made, Ottr.me! Inspired by Jet Blue's massive and extensive data set, we decided to create a multifaceted travel website at YHack 2015. Using the data set from Jet Blue, we were able to make a "Go Far" section that generated the user potential vacation packages they may be interested in based on their previously selected preferences. The site calls the Priceline API to recommend hotel deals alongside airfare from Jet Blue. The site additionally includes a "Go Local" function that displays local activities that may be of interest, drawn from the Yelp, Fandango, and Stubhub APIs, based on user prefences and zip code. Check out our Devpost page or use it yourself!

Real Drone

I travelled out to UC Berkeley with Eric Liu and Zamir Johl last fall to build Real Drone, an Intel Real Sense-controlled custom flight controller for my DJI Phantom quadrotor. Wanting to work on a reverse engineering hack, we disassembled the Phantom controller to figure out how the physical controls interacted with the transmitter. We were able to disconnect the potentiometers on the joysticks and replace them with digial potentiometers, which we controlled with SPI (Serial Peripheral Interface) using an Arduino. We fed data to the Arduino using a C++ program that gathered hand movement data from the Real Sense and translated it to flight directions. Check out our Devpost pagehere or watch the video.

Project Horus

At my second hackathon of the season, I teamed up with Eric Liu and Max Connor to head down to New York City for the weekend, where we decided to work with local startup Clarifai's API to build Project-Horus, a web application intended to make language education simpler and more effective by translating objects from photographs rather than direct vocabulary translations. Project-Horus utilizes Clarifai, Yandex.translate, and a Javascript/HTML5 selection tool that we created from scratch at HackNY. After a two minute pitch and demonstration for the HackNY judges, we won third place overall and “Best Use of Amazon Web Services” from Amazon. Check out our Devpost page or Read More.

Piano Lights

Along with Max Connor, Kurt Jaisle, and Zach Marcus, I designed, constructed, and programmed a hardware hack to simplify learning to play piano. We utilized a 3D printer to design and manufacture the device, Arduino to operate an array of LEDs and to receive MIDI input, and Java to create an application called Song Maker, allowing Piano Lights users to create their own songs, which can then be shared online. Our final design utilizes a TFT touch screen and MIDI input to easily interact with users of all ages and have an awareness of what note(s) a user is playing in order to check for correctness. Check out out Devpost page or Read More.

Smart Station

At my first undergraduate hackathon, I developed a hardware hack called Smart Station to provide users with an interactive sensor bank that is portable and communicates via text message. Smart Station runs using an Arduino, keypad, LCD display, and an array of seven sensors. A micro SD card shield mounted on the back of the LCD (shown below) allows users to store data sets at user-defined intervals for up to one week at a time. The device recognizes and responds to SMS requests like "humidity" and "text me if temperature is greater than 120". Check out our Devpost page or Read More.

IoC (Internet of Chess)

At Harvard's third hackathon, I worked with Kevin Mcdonough to build an Internet of Things-enabled ChessBot. The chess robot was capable of using openCV to analyze real-world moves and play against a present human, however it was also capable of taking in moves from a web interface and making those moves physically on the board. I was responsible for the entire hardware design and construction of this project, which involved setting up a Raspberry Pi camera to analyze moves and sending the changes to the ATMega328 running the stepper drivers. The Pi 3 (wifi enabled) also retrieved changes from Kevin's API and sent changes to the Arduino, which then translated them to coordinates and made the physical moves.