Hi, I’m Amir

I’m finishing an M.S. in Data Science at Columbia University (December 2026), with a background in financial economics. I’m interested in machine learning for defense and in quantitative finance. Check out some of my projects below!

Selected work

PhishGuard

A client-side Chrome extension that scores Gmail messages for phishing, with nothing leaving the browser.

JavaScript
Python
LightGBM
TF-IDF
Logistic Regression
Chrome Manifest V3

Overview Repo

Part of the motif graph linking statistics concepts such as the EM algorithm, log-likelihood and maximum likelihood

Structural Recall

A spaced-repetition system where FSRS schedules reviews and a motif graph queues related cards after one is forgotten.

Python
FSRS-6

Overview

What I'm working on

Screenshots sorted into code, figure and text columns with the model's confidence on each, including one misclassified flowchart

Screenshot Topic Classifier

A three-class screenshot classifier (code, figure, text): a linear probe on frozen CLIP image embeddings, calibrated confidence routing with blind human review, and leave-one-origin-out evaluation.

Python
CLIP
scikit-learn
SQLite

Overview

Other projects

3D scatter of three classes with the first three principal components drawn as axes

PCA visual explainer

An interactive 3D explainer for principal component analysis.

JavaScript
Three.js
WebGL

Overview

Tiles on the faces of a cube seen through a fisheye lens, with one column lit yellow and a white diamond marking the player

Super Cube

Play my game!

A rhythm survival game on the surface of a cube.