Current
xAI — Grok Imagine
Member of program staff at xAI, working on human data for image and video generation models powering Grok Imagine.
Artist & Technologist
Current
Member of program staff at xAI, working on human data for image and video generation models powering Grok Imagine.
Research
Pioneering work in jailbreaking, prompt extraction, and adversarial techniques for large language models.
First-to-market AI tools and growth experiments, built at the frontier before the platforms caught up.
2023
The first Stable Diffusion images generated inside ChatGPT — months before DALL-E 3 integration. Exploited the model's ability to render markdown images from external URLs, routing descriptions to a Stable Diffusion API that returned generated images directly in the chat window.
2023
Extracted ChatGPT's system prompt when it first acquired browsing capabilities, then used the same format to define custom "functions" — small commands that steered the model's writing style. Surprisingly effective steering with minimal prompting.
2023
Automated product review system. Paste an Amazon link — it scrapes the most helpful reviews, summarizes via GPT-3.5/4, generates a comprehensive review highlighting all pros and cons, compares multiple products, and auto-publishes to WordPress via REST API. Amazon built something similar a year later.
2023
Built a pipeline that pulls images from Instagram, captions them with vision models, generates carousel code for WordPress, then creates Pinterest-optimized captions and auto-pins to the right boards. Grew a test account to nearly 100M impressions in six months.
Film
Director and visual artist. Matte paintings, VFX, and films that bridge art and technology.