Index
About
Background
Fifteen years in AI research, applied AI, product management, strategy consulting, and data science. Prior organizations include IBM, Stability AI, Petuum, Dell, and the World Bank. Based in Austin, Texas.
AI Model Benchmarking
StoryBench · DharmaBench
Model evaluations run on a 2× NVIDIA DGX Spark cluster. StoryBench measures creative-writing capability. DharmaBench measures a model's unprompted philosophical proclivity across 25 tradition judges.
LLM Finetuning
Dharmatune
Finetuning open models to hold a specified worldview without loss of capability. Current results: a Gemma 4 12B ranked first of 25 traditions on the Catholic worldview (85/100), and a second model ranked first on the Mahayana Buddhist worldview (58/100).
Writing
Signal · Cyber · Articles
Signal is the daily AI research briefing. Cyber is the daily briefing on agentic threats and supply-chain compromises. Articles are longer pieces on agentic systems, context engineering, memory architectures, and harness design.
Teaching
Workshops · Austin
Free in-person AI workshops in Austin. Hands-on sessions for builders, founders, product managers, and general practitioners.
Apps & Projects
Thinkybooks · Prompt Engines · Video Terminal
Shipped applications. Thinkybooks: storybooks for kids. Prompt Engines: agent orchestration systems. Video Terminal: node-based video generation. Additional educational games and tools.
Latest
Nov 16
Daily Cyber Briefing — November 16, 2026
Cyber
Nov 15
Daily Cyber Briefing — November 15, 2026
Cyber
Nov 14
Daily Cyber Briefing — November 14, 2026
Cyber
Nov 13
Daily Cyber Briefing — November 13, 2026
Cyber
Nov 12
Daily Cyber Briefing — November 12, 2026
Cyber
Nov 11
Daily Cyber Briefing — November 11, 2026
Cyber
Nov 10
Daily Cyber Briefing — November 10, 2026
Cyber
Nov 6
Daily Cyber Briefing — November 6, 2026
Cyber