Hi! I'm Azmi, a .NET engineer from Indonesia.
I've been building AI-augmented development workflows for the past year — running a full self-hosted stack (Qdrant + Ollama + n8n + custom .NET RAG gateway) on a Proxmox homelab. Most of my work is in private enterprise repos (oil & gas e-procurement for PetroChina), but I open-source my AI tooling where I can.
One of my recent projects: RTK — a token compression hook for Claude Code that intercepts shell commands and saves 60–99% tokens on repetitive ops like builds and git. Also built a RAG capture pipeline that structures coding sessions into a queryable Qdrant knowledge base.
I found linux.do through threads shared externally, and the depth of discussion — especially around LLM APIs, self-hosted infra, and AI tooling — is unlike anything I've seen on English-language forums.
Would really appreciate an invite to learn from and contribute to this community.
GitHub: https://github.com/figulazmi
Email: azmi.codes@gmail.com
Thank you! 🙏
Hi! I'm Azmi, a .NET engineer from Indonesia.
I've been building AI-augmented development workflows for the past year — running a full self-hosted stack (Qdrant + Ollama + n8n + custom .NET RAG gateway) on a Proxmox homelab. Most of my work is in private enterprise repos (oil & gas e-procurement for PetroChina), but I open-source my AI tooling where I can.
One of my recent projects: RTK — a token compression hook for Claude Code that intercepts shell commands and saves 60–99% tokens on repetitive ops like builds and git. Also built a RAG capture pipeline that structures coding sessions into a queryable Qdrant knowledge base.
I found linux.do through threads shared externally, and the depth of discussion — especially around LLM APIs, self-hosted infra, and AI tooling — is unlike anything I've seen on English-language forums.
Would really appreciate an invite to learn from and contribute to this community.
GitHub: https://github.com/figulazmi
Email: azmi.codes@gmail.com
Thank you! 🙏