Zachary Zusin

Zachary Zusin

About Me

I am a recent Computer Science graduate from Columbia University with a strong foundation in full-stack development, data science, and machine learning. I’ve held several machine learning research and software engineering roles and have experience applying my knowledge in software and statistics to domains ranging from computational neuroscience to automatic speech recognition. Please feel free to reach out with any opportunities or simply to connect!

Languages

Front and Back End

Data Analysis and Machine Learning

Projects

Home Cooking Helper

Home Cooking Helper

This web application serves as a comprehensive platform for home cooking, where users can receive guidance on anything they would need to know to prepare food in the comfort of their own home.

View on GitHub
PythonPostgreSQLFlaskHTMLCSSJavaScript
Factory Workflow Tracker

Factory Workflow Tracker

A lightweight Manufacturing Execution System (MES)-inspired web app simulating job orchestration and monitoring in a factory environment.

View on GitHub
ReactPostgreSQLFastAPITypeScriptPython
Bird, Dog, or Reptile?

Bird, Dog, or Reptile?

This research project proposes and implements a hierarchical image classification model using a two-headed ResNet-50 architecture to classify animal images into both coarse (bird, dog, reptile) and fine-grained (chihuahua, rooster) categories, while handling unseen sub-classes via threshold-based novelty detection.

View on GitHub
PyTorchPythonMachine LearningComputer Vision
Action Segmentation and Behavioral Analysis in Mice

Action Segmentation and Behavioral Analysis in Mice

This is the code I wrote for my Computational Neuroscience Research internship in which I developed diagnostics to assess the efficacy of Temporal Convolutional Network (TCN) models in predicting behavioral states of mice.

View on GitHub
PythonMachine LearningData Visualization
TDNN-Conformer

TDNN-Conformer

This research paper proposes and implements a modification to the Conformer model for the task of automatic speech recognition.

View on GitHub
PythonMachine LearningSpeech Recognition
Wikipedia Article Clustering

Wikipedia Article Clustering

This project fetches and preprocesses Wikipedia articles, converts the text into semantic embeddings using Sentence-BERT, and then applies UMAP for dimensionality reduction and K-Means clustering. It then creates an interactive scatter plot that visualizes the clustered articles, allowing users to explore their relationships based on semantic similarity.

View on GitHub
PythonNatural Language ProcessingScikit-Learn
Bitcoin Price Tracker

Bitcoin Price Tracker

This code implements a real-time Bitcoin price tracking application built with Elixir and Phoenix, featuring LiveView for real-time updates and PostgreSQL for price history storage.

View on GitHub
ElixirPhoenixPhoenix LiveviewPostgreSQL
Exploring the Determinants of Life Expectancy

Exploring the Determinants of Life Expectancy

This project develops a statistical model to examine the relationship between life expectancy and various social, economic, political, and geographic variables.

View on GitHub
RLinear Regression Models
State Space Model POS Tagger

State Space Model POS Tagger

This project implements a state space model-based Part-of-Speech tagger using MATLAB.

View on GitHub
MATLABNLP
Wonderful Workouts

Wonderful Workouts

This web application serves as a platform to teach individuals the basics of weight training, allowing users to learn new exercises and create their own custom workouts.

View on GitHub
HTMLCSSJavaScript
Neural Network Dependency Parser

Neural Network Dependency Parser

This code implements a neural network-based dependency parser implemented in PyTorch that predicts transitions for an arc-standard dependency parser.

View on GitHub
PythonPyTorchNatural Language Processing
CKY Parser

CKY Parser

This project implements the Cocke-Kasami-Younger (CKY) algorithm for parsing Context-Free Grammars (CFG) and Probabilistic Context-Free Grammars (PCFG).

View on GitHub
PythonNatural Language Processing
Trigram Language Model

Trigram Language Model

This project implements a trigram language model in Python for text analysis and essay classification.

View on GitHub
PythonNatural Language Processing
Personal Portfolio

Personal Portfolio

This website! I designed and created this site from scratch using Next.js, Tailwind CSS, and Framer Motion.

View on GitHub
Next.jsTailwind CSSFramer Motion