No fluff. Fullstack, system design, and AI — explained precisely
Fullstack Software Engineer specialising in cloud architecture, system design, and AI. I write deep dives on building production systems — the decisions, the tradeoffs, and the reasoning behind them. No fluff, no tutorials you've seen a hundred times.
Markdown supported · ⌘↵ to post
Kartik N V J K
Least-privilege tools per task plus an LLM-as-judge inside the executor is a strong pairing, you bound what each step can touch and score its output before it counts as done. The recovery detail, resetting IN_PROGRESS to PENDING on restart, is what people skip until a crash mid-task wedges the whole plan. For the judge, is it scoring each task against the plan's stated intent, and can a failed score reopen the task or does it only flag for the human A/M/R loop?