Commit

I've been building Amaru, an engineering execution platform for the age of AI agents. The idea started with a question: as coding agents become increasingly capable, what do we need to make the engineering work around them reliable? Writing code is only one part of the process. Real engineering work involves dependencies, workflows, tools, artifacts, execution environments, interruptions, and verification. Amaru explores a different layer of the stack: coordinating that work without replacing the agents or tools that perform it. I'm interested in making engineering workflows explicit, execution state durable, progress observable, and recovery a first-class concern. The goal isn't another AI coding assistant. It's infrastructure for orchestrating engineering work across humans, agents, and tools while keeping engineers in control. I'm still exploring the architecture and working through the trade-offs, and that's part of what makes this project interesting to me. I'd love to connect with other developers building agent infrastructure, orchestration systems, or tools for reliable AI-assisted engineering. What are you building in this space?