Ran Qwen 3.5 27B via Ollama as a persistent background agent for 30 days. Not a demo. Honest results.

Build a robust framework for local persistent agents that manages state, memory, and task scheduling outside of a chat interface. This fills the gap between raw LLM inference and a reliable, long-running background worker system.

Suggested repo: AgentCron

"Turn any local LLM into a persistent, scheduled worker for your OS."

Estimated effort: 40h