Exploring the foundations of AGI
Archemind is an interest collective of engineers and researchers working on AI agents. Our members build and study agents inside internet companies and AI startups across Silicon Valley and China, and gather here as individuals.
We care about the machinery that makes language models act: LLM harnesses, agent scaffolding, context engineering, evaluation, tool use, memory, and multi-agent orchestration. We publish research notes and run experiments in the open.