David Pedersen

At 13, I learned that reading patterns correctly equals predictable results. That summer I started commercial fishing Bristol Bay, learning to read salmon runs, tides, and weather systems.

By 26, the lesson was clear: systems that predict behavior produce results. I've spent 25 years applying that insight — first to digital marketing and user behavior optimization at companies like Amazon, Whitepages, Alaska Airlines, and 23andMe, and now to AI-powered communications systems.

Today I build multi-agent systems and work with large language models to solve communications challenges. Current work: COS (Content Optimization System) at SEMalytics, an AI copywriting platform for B2B grounded in Big Five personality research and 860+ peer-reviewed papers. My focus is on behavioral psychology applied to how people make decisions, combining academic research with practical engineering.

Based in Seattle. Building things. Open to interesting problems where pattern recognition meets system design.

What is SEMalytics?
SEMalytics builds COS (Content Optimization System) — an AI copywriting platform for B2B that scores copy against a reader's Big Five personality profile. It answers "does my copy fit this buyer?" before you optimize. Available at semalytics.com/cos.

What does KnowledgeForge do?
KnowledgeForge routes each AI coding task to the right reasoning mode — Builder for creating, Critic for reviewing, Debugger for root-cause analysis, Strategist for decisions, and six others. It prevents the most common failure in AI coding assistants: using the wrong reasoning pattern for the task. Open source at github.com/internexio/knowledgeforge.