Cutting LLM inference cost and latency without making the feature worseThe first bill for an LLM feature is usually a surprise in one of two directions. Either it is tiny and nobody thinks about it again until usage grows, or it is larger than expected and someone asks f25KAA1d ago · 9 min readmahmadmanzoor.hashnode.dev
Top 10 skills a software engineer needs in 2027Most of the code I ship now is typed by an agent. I write the spec, Claude Code writes the diff, the checks run, and I read what came back. That changed which of my skills earn their keep. Typing spee2d ago · 13 min readmahmadmanzoor.hashnode.dev
Caveman mode for Claude Code: fewer output tokens, same factsClaude Code is polite. Ask why a component re-renders and you get a paragraph that restates your question, explains the cause, suggests a fix and offers more help. In a terminal, at the end of a long 3d ago · 10 min readmahmadmanzoor.hashnode.dev
Harness engineering: the skill that took over from prompt engineeringIn 2024 the skill everyone talked about was prompting: find the right words and the model does the right thing. In 2025 it was context: hand the model the right files and docs, and the exact wording m27AKA4d ago · 18 min readmahmadmanzoor.hashnode.dev
Setting up REA for agent-driven reverse engineering in Claude CodeSometimes the answer is not in a source repository. It is in a build: the Electron app that went out last month, an APK on a test phone, a native binary from an old release with no tag. Questions like4d ago · 10 min readmahmadmanzoor.hashnode.dev
