# Ethan Austin-Cruse

> I ship software from kernel to interface.

- Status: Seeking Summer 2027 opportunities
- Location: Atlanta, GA
- Email: contact@ethanaustincruse.com
- Website: https://ethanaustincruse.com
- GitHub: https://github.com/ethandom
- LinkedIn: https://www.linkedin.com/in/ethan-austin-cruse-68221a2a9
- Resume (PDF): https://ethanaustincruse.com/resume.pdf

## Education

- Georgia Institute of Technology: B.S. Computer Science, minor in Mathematics
- Expected graduation: May 2028
- Coursework: Computer Organization & Programming, Data Structures & Algorithms, Object-Oriented Programming, Linear Algebra, Multivariable Calculus, Probability & Statistics

## Skills

- Languages: TypeScript, JavaScript, Java, Python, C, SQL, HTML/CSS
- Frameworks: React, React Native, Node.js, Hono, Spring Boot, Tailwind
- Tools: Git, Linux, PostgreSQL, Docker, Vercel, Supabase, CI/CD, Claude Code, Codex
- Libraries: Zod, AI SDK, Playwright, Pandas, TensorFlow, scikit-learn, Beautiful Soup, Radix

## Away from the keyboard

- Gold Medal, WFDF World Ultimate Championships, USA Ultimate National Team (Jun 2026). Defensive specialist, Logroño, Spain.
- Eagle Scout, Scouting America (Aug 2025). Led 8 youth volunteers through 75 hours of work building mushroom inoculation tables to increase local produce yield.

## Work history

## Founder, Conturo

- Organization: Conturo
- Role: Founder
- Period: Feb 2026 – Present
- Stack: TypeScript, Python, Hono, PostgreSQL, Deno, MCP, Claude Code, Codex
- Project write-up: https://ethanaustincruse.com/projects/conturo.md

### Highlights

- Shipped a REST API that lets nonprofits move 100k+ contacts between conturo.io and their CRM, email and analytics tools.
- Keeps a B2C database current by processing 1M+ contact datums a month with cron jobs, Playwright and the OpenAI API.
- Built an MCP server and chat UI so nonprofits can request contact exports in natural language.
- Wrote a crash-safe export worker (atomic job claiming, lease-based retry) covered by 200+ Deno tests.

## Apprentice to Google VP

- Organization: Google
- Focus: Software and machine learning
- Type: apprenticeship
- Period: Jan 2025 – Feb 2025
- Stack: Java, Linear Algebra, Machine Learning
- Project write-up: https://ethanaustincruse.com/projects/chess-engine.md

### Highlights

- Designed a chess engine with alpha-beta pruning that reaches 500k–650k evaluations per second, used to benchmark a reinforcement-learning agent.
- Researched how gradient descent and cost functions are applied in LLMs.

## Data Engineer, Center for Puppetry Arts

- Organization: Center for Puppetry Arts
- Role: Data Engineer
- Period: Nov 2024 – Nov 2025
- Location: Atlanta, GA
- Stack: Python, Playwright, Beautiful Soup, HTML, ChatGPT

### Highlights

- Aggregated 150k+ marketing data points with Python scripts, helping grow revenue from a new education segment 5x.
- Made collecting education information up to 100x faster with Playwright and AI parsing, reaching 30,000+ new educators.

## Apprentice to Merrill Partner

- Organization: Merrill Lynch
- Focus: Financial operations research
- Type: apprenticeship
- Period: Jan 2024 – Feb 2024

### Highlights

- Built an analysis portfolio of Merrill's internal operations, learning from a principal partner, financial advisors and security officers.
- Presented the findings to mentors.

## Projects

## Conturo

> Contact sync for nonprofits across their CRM, email and analytics tools.

- Period: Feb 2026 – Present
- Role: Founder, sole technical founder
- Status: active
- Stack: TypeScript, Hono, Zod, PostgreSQL, Deno, Playwright, OpenAI API, MCP
- Live: https://conturo.io
- Related work history: https://ethanaustincruse.com/experience/conturo.md

### Highlights

- Shipped a REST API (Hono, Zod, PostgreSQL) that lets nonprofits move 100k+ contacts, each with 10–20 identity fields, between conturo.io and their CRM, email and analytics tools.
- Processes 1M+ contact datums a month with cron jobs, Playwright and the OpenAI API to keep a B2C database current.
- Built an MCP server and chat interface so nonprofits can request exports in natural language.
- Wrote a crash-safe export worker with atomic job claiming and lease-based retry, covered by 200+ Deno tests.

### Metrics

| Metric | Value |
| --- | --- |
| Contacts per transfer | 100k+ |
| Fields per contact | 10–20 |
| Datums / month | 1M+ |
| Worker tests | 200+ |

## USB Relay Kernel Driver

> A Linux kernel driver and CLI for USB relay boards.

- Period: Oct 2025 – Dec 2025
- Status: shipped
- Stack: C, Linux Kernel, USB, FTDI, Make
- Repository: https://github.com/ethandom/SainSmart5VUSBRelayDriver

### Highlights

- Kernel module that binds to the board's FTDI FT232R chip and exposes each board as /dev/usbrelayN, with a one-byte mask ABI that drives all 8 channels.
- Enables FTDI bit-bang mode with vendor USB control requests, so all 256 output states fit in a single byte write.
- Handles hot-plug through probe/disconnect, endpoint discovery, and bulk transfers with 1 s timeouts.
- Supports up to 4 boards at once, with IDA minor allocation, a mutex per device, and leak-free cleanup on every error path.
- Includes relayctl, a user-space CLI that turns commands like "set 1 on" into bit masks, with a functional test script.

### Metrics

| Metric | Value | Note |
| --- | --- | --- |
| Channels | 8 | per board |
| Boards | 4 | max concurrent |
| States / byte | 256 |  |
| Transfer timeout | 1 s |  |

## Tempr

> Swipe-to-discover music, built on Spotify.

- Period: Feb 2026
- Role: Owned swipe discovery, team of 4
- Status: shipped
- Stack: TypeScript, React Native, Supabase, Gemini API, Spotify API
- Repository: https://github.com/ethandom/TemprHackalytics

### Highlights

- I placed 2nd in the GT Hacklytics Entertainment track, shipping a React Native app with Spotify sign-in in 36 hours.
- I built a Tinder-style, short-form music discovery feature as the base for a vector-similarity recommendation engine over Spotify song vectors.
- The team built multimodal Gemini pipelines that turn a text prompt or photo into a playlist through the Spotify API.

### Metrics

| Metric | Value | Note |
| --- | --- | --- |
| Placement | 2nd | Hacklytics, Entertainment track |
| Build time | 36 h |  |
| Team | 4 |  |

## Chess Engine

> An alpha-beta chess engine, built at Google to benchmark a reinforcement-learning agent.

- Period: Jan 2025 – Feb 2025
- Role: Software Engineering Apprentice, Google
- Status: archived
- Related work history: https://ethanaustincruse.com/experience/google.md

### Highlights

- Designed a chess engine with alpha-beta pruning that reaches 500k–650k evaluations per second, used to benchmark a reinforcement-learning agent.

### Metrics

| Metric | Value |
| --- | --- |
| Evals / sec | 500k–650k |