LLM 튜토리얼

Set up vLLM on a H100 GPU, serve Qwen3.5-27B, connect OpenCode, and test fast agentic coding with long context support.

Learn how to fine-tune Qwen3.5-9B on the Latex OCR dataset using QLoRA in Unsloth Studio, then test the model in chat and export it in GGUF format for local inference.

Learn how to use NVIDIA’s Nemotron 3 Super through OpenRouter to analyze a GitHub repository without RAG, using prompt packing and a Gradio app.

Build reliable AI apps faster by turning ad-hoc prompt checks into structured LLM evaluations with Promptfoo, from local test suites to automated CI.

Learn how to design and maintain a lean CLAUDE.md file, so Claude Code reliably follows your project’s rules, conventions, and workflows in every session.

Learn how to use Qwen 3.5 small models locally through Ollama to turn a short gameplay clip into a playable HTML game.

Discover how Google's newest specialized reasoning model can accelerate your data science workflows, interpret complex datasets, and write robust code.

Break your AI coding agent out of its sandbox. Discover how to use Codex CLI and MCP to connect LLMs to live data, external tools, and multi-agent workflows.

Learn how Cursor rules guide AI coding, from defining project-level context and file-matching patterns to maintaining accurate, up-to-date rules over time.

Learn Claude Code best practices from production teams. Use plan mode, CLAUDE.md files, and test-driven development to make the most of Claude's context window.

Learn how to use Claude Code’s Plan Mode to explore your codebase in read-only mode, surface questions, and review structured plans for multi-file changes.

Discover how to use Claude Code Router to connect to multiple providers, use the optimal LLM for every task, and reduce API cost without changing your workflow.