Jeremy Kamber
Full-stack Developer and Product Manager based in Seattle. I build AI products with a focus on making LLMs feel more human.
jnk-skills: A Development Workflow for AI Agents
I published the workflow I use to build software with AI agents. It's organized into beats: pickup, understand, decide, design, implement, verify, debrief. Each beat ends with a gate. You stay the pilot. The agent does one job at a time. The point is that every session leaves better software and better understanding. One without the other is incomplete. The repo also includes teach, a tutoring skill based on learning science.
Get the skills on GitHub→Kynd: AI-Powered User Testing with Synthetic Personas
Generate realistic synthetic personas from minimal input and run them against live websites for automated user testing, pricing analysis, and behavioral research. Built with Hexagonal Architecture and six research-backed inference-time techniques.
Read the case study→Strata: A Tiered Memory System for AI Agents
A zero-dependency tiered memory system built for AI agents. Separates algorithmic lifecycle triggers from LLM compression so you only pay for intelligence when it's needed. 137+ tests, 3-tier architecture, filesystem-first design.
Read the case study→Don't Vibe Code. Engineer Code.
My framework for engineering with AI agents. Writing syntax is a solved commodity. Communicating intent and defining requirements is where the real engineering is.
Read the article→