Hi, I'm Aadi!
Student Developer
- Passionate about AI safety, ML robustness, and adversarial ML
- Experienced in front-end & back-end development and proficient with algorithmic programming
Contact Me
About Me
My journey as a prospective data scientist
Student developer currently prototyping assistive technology to aid the impaired. Lead Writer of the Data Driven Digest, a student-run data science blog about its impact on the populations of today.
experience
projects
organizations
Skills
My technical levelFront-End Development
More than 4 yearsHTML/CSS
JavaScript
TypeScript
Back-End Development
More than 3 yearsJava
Python
C++
Data Analysis
More than 2 yearsR
MATLAB
NumPy
Machine Learning
More than 2 yearsScikit-learn
TensorFlow
PyTorch
Qualifications
Everything that got me hereHigh School Student
Santa Clara High SchoolStudent & Mentor
STEM Leadership InstituteNeural Networks
MIT LL BWSIData Analysis
UC Berkeley ATDPCo-founder/App Dev
NaNalysis (Hack+ Project)Student Mentor
Zenith (Hack+ Project)Founder/Lead Software Dev
Overhead InitiativeAdversarial ML/Game Theory Researcher
Pioneer Academics Research ProgramPresented at State of the Community 2023
Project InventGold Qualification
USACO2x PVSA Gold
Zenith (Hack+ Project)4x AIME Qualification
MAAExplore Award &
$500 Seed Grant + $500 Grant From SLI
Project Invent
Portfolio
A few examples of recent projects
re-paragrapher
Browser extension that uses natural language processing principles to "reparagraph" onscreen text upon click. Helps make blocky text more readable.
In Development
Overhead Movement Alarm Radar
Prototype using ultrasonic sensors and vibration motors to help the visually impaired detect overhead obstacles and their direction via haptic feedback.
In Development
Cog*Ify
App using a CNN trained on NCS song spectrograms and ‘most replayed’ data from YouTube to predict the likely ‘most popular’ portion of each song in a Spotify playlist and order and mix them with cosine similarity and crossfade.
Check it out!
Personal Site
Well designed, clean and professional personal website with portfolio created using HTML, CSS, and JavaScript, and without any frameworks. Has options for both light and dark mode.
Check it out!
dig-interpret
Device using a convolutional neural network and text-to-speech to help the visually impaired audibly perceive numerical data in their environment. Overhauled from October to December 2021 to interpret all text using optical character recognition; more info to be published at a later date.
Check it out!