Brain in the Loop
Exploring the edges of Artificial Intelligence.
Deep dives into architecture, alignment, and agents.
On-Device SLM Investigation: Prompt Iteration, Hallucinations, and Agentic Tool Use
Four 3B-class on-device language models, two HealthKit-style tasks, and the failure modes that matter when an answer is going to be shown to a person trying to understand their own body.
Governing Claude: Enforcing Senior Developer Standards on an LLM
How I treated Claude like a reckless Junior Developer to build a production-grade iOS health app with strict physics and medical standards.
The Rise of the 'AI Intern': How We Built a Multi-Agent Strategy Engine Without a Team of Software Engineers
Moving from data gathering to strategic thinking in Pharma Portfolio Management using Dataiku, Gemini 3.0, and Agentic workflows.
The Economics of Building a 128,000 LOC iPhone App with AI
A transparent financial breakdown of building Ahimo, a production-grade iOS health app, using AI-driven development - achieving 97-99% cost reduction compared to traditional development.
When an AI Invents Market Wisdom
A case study of how Google Gemini invented a convincing but nonexistent market aphorism under geopolitical pressure