Insights &
Thoughts
Exploring the frontiers of software architecture, AI agents, and modern development practices.
Beating llama.cpp from Scratch on Consumer AMD: Building Strata, a Native Rust + HIP Local AI Engine
Software Architect Mihai Farcas details engineering Strata: an ultra-fast local LLM inference engine for Qwen 3.5 on AMD Radeon RX 7900 XTX (gfx1100). How custom HIP kernels, split-KV reduction, and zero-allocation Rust beat llama.cpp and Ollama at 83.0 tok/s BF16 decode.
n8n + Argo = ❤️: Offloading CPU-Heavy Steps to Argo Workflows
Offload CPU-heavy n8n steps to Argo Workflows on Kubernetes. n8n keeps event routing, webhooks, and visual business logic while Argo runs the compute, so no job blocks a worker slot. Lab data from a k3s cluster.
Repurposing an i7 Chromebox into a Proxmox VE Server: MrChromebox UEFI & Stability Guide
Why I am wiping my homelab to rebuild from scratch, how to flash the latest MrChromebox UEFI Full ROM on an Intel Core i7 Chromebox, and the mandatory kernel stability flags needed to run consumer mobile silicon 24/7 without crashing.
ArgoCD
GitOps continuous delivery tool for Kubernetes, adopted over Flux.
AWS
Amazon Web Services, on hold in favor of GCP.
Azure
Microsoft Azure, rejected in favor of GCP.
Black
The uncompromising Python code formatter that eliminated style debates, now superseded and rejected in favor of Ruff.
Celery
Distributed async task queue for Python, currently being evaluated.
Claude Code
Agentic coding CLI/IDE assistant from Anthropic.
Showing 1 to 9 of 68 items