jonathan-wrenn › src › about.tsx

Jonathan Wrenn

// Building AI systems that ship.

import type { Engineer } from "@columbia/cs-2026";

const jonathan: Engineer = {
  role:       "Incoming SWE @ Rasa Legal",
  school:     "Columbia University · CS '26",
  building:   [
    "AI pipelines",
    "data tools",
    "automation systems",
  ],
  background: [
    "Rasa Legal — automated 10k+ legal records",
    "Princeton — mined 151M quantum structures",
    "Justice Through Code — teaching CS to returning citizens",
  ],
  contact:    "jw4544@columbia.edu",
};

export default jonathan;

See my work Resume ↗

// GitHub activity — github.com/jonowrenn

jonathan-wrenn › src › experience.tsx

// Work experience & education

Experience

Incoming · Jul '26 /01

Remote /02

Wrote Python scripts leveraging LLMs to automate document classification and analysis across 10k+ legal records.
Developed and optimized data classification pipelines querying large-scale PostgreSQL databases.
Conducted data quality audits detecting duplicates, taxonomy drift, and label conflicts.

Princeton, NJ /03

Filtered and analyzed 151 million Quantum Cascade Laser structures using Python data mining workflows.
Built tools reducing analysis time from 5+ minutes to under 30 seconds.
Validated optimized QCL designs for sensing and autonomous vehicle applications.

New York, NY /04

Hold weekly office hours helping students complete assignments and debug AI and ML projects across a Python-based curriculum.
Teaching programming and computer science fundamentals to formerly incarcerated individuals transitioning back into society.

Remote /05

Designed AI-driven recruiting workflows reducing hiring cycles from 2–4 weeks to under 10 days.
Built an AI chatbot serving 600+ monthly users at under $5/month operating cost.
Automated video interview screening, eliminating 20+ hours of weekly manual review.

// Education

New York, NY

Columbia University

B.S. Computer Science · Class of 2026

Activities: Club Basketball · Justice Through Code TA · Columbia Change Agent Award (2025)

jonathan-wrenn › src › projects.tsx

// Personal projects — things I built because I wanted them to exist

Projects

Machine Learning · Finance Python

ml-stock-predictor

LSTM neural network that predicts AAPL stock prices using TensorFlow/Keras and yfinance market data — built to explore time-series forecasting and deep learning.

PythonTensorFlowKerasLSTMyfinance

GitHub ↗ Live ↗

AI Agent · Full-Stack TypeScript

fashion-agent

AI-powered wardrobe inventory that catalogs your clothes and generates outfit ideas with gap analysis — built as a full agentic workflow with TypeScript.

TypeScriptAI AgentNext.js

GitHub ↗ Live ↗

AI & Automation · macOS Python

screenshot-to-ai

macOS menu bar app that watches your clipboard and auto-pastes screenshots directly into Claude or ChatGPT — zero friction from capture to AI response.

PythonmacOSClaude APIChatGPTClipboard

GitHub ↗ Live ↗

AI Tool · Education Python

cs-study-assistant

Multi-turn GPT chatbot with 6 expert modes (Code Review, ML Theory, Interview Prep, etc.), conversation memory, BYOK support, and deployed live on Hugging Face Spaces.

PythonOpenAI SDKGradioHugging Face

GitHub ↗ Live ↗

Full-Stack · Productivity TypeScript

interview-prep-tracker

Next.js + TypeScript app for tracking LeetCode interview prep — log problems, track progress by category, and stay accountable through a clean dashboard UI.

Next.jsTypeScriptReact

GitHub ↗ Live ↗

Web App · JavaScript JavaScript

weather-app

Real-time weather app with a 7-day forecast, city search, and geolocation support — clean UI pulling live data from a weather API.

JavaScriptWeather APIGeolocation

GitHub ↗ Live ↗

Web App · Education Python · HTML

muscle-map

Interactive anatomy learning app with a clickable SVG body map — click any muscle group to get detailed info. Built with Flask and custom SVG path mapping.

PythonFlaskSVGHTML/CSS

GitHub ↗ Live ↗

jonathan-wrenn › src › skills.tsx

// Tools & technologies