Cost Optimized Gpu Cluster Scaling

via “resource budgeting and cost optimization for gpu experiments”

ARIS ⚔️ (Auto-Research-In-Sleep) — Lightweight Markdown-only skills for autonomous ML research: cross-model review loops, idea discovery, and experiment automation. No framework, no lock-in — works with Claude Code, Codex, OpenClaw, or any LLM agent.

Unique: Implements cost-aware experiment orchestration with pre-execution cost estimation, budget enforcement, and cost-per-paper metrics. Enables cost-optimized experiment selection (greedy algorithm to maximize value within budget). Most research tools ignore costs; ARIS makes cost optimization a first-class concern.