Catalog

Fine-tuning: when (and when not)

A decision tree before you spend a dollar on GPUs.

16 min 2,840

Agents — from prompt to production

A 12-step path from "it works on my laptop" to live traffic.

110 min 4,910

Shipping safe LLM features

Red-teaming, jailbreak budgets, and rollback playbooks.

35 min 3,120

Prompt engineering without tears

Eight patterns that survive model upgrades.

22 min 22,840

Embeddings — the quiet workhorse

What they are, why they matter, and 6 ways to use them.

30 min 11,430

Building evals you actually trust

From vibe-checks to LLM-as-judge to gold sets.

50 min 5,210

Model Context Protocol from first principles

Why MCP, what it replaces, and a server in 30 lines.

60 min 9,120

RAG in 2026: still worth it?

When long-context wins. When retrieval wins. When neither.

18 min 7,320

GPT Realtime voice — the practical handbook

Build a low-latency voice agent that interrupts politely.

95 min 6,210

Gemini 2M-token context: when long actually helps

Benchmarks, tradeoffs, and 4 patterns that actually win.

28 min 4,830

How to use Claude tool-use in 5 steps

Wire up your first tool-using agent in under an hour.

42 min 12,480

Your first prompt in five minutes

No setup. Just type. See output.

5 min 38,420