build-aiewf-2024-rag

Build AIEWF 2024 RAG stacks

Action playbook from eight AI Engineer / World's Fair talks. Do not summarize talks — pick a workflow and execute it.

Supporting files (read when needed):

• workflows.md — workflows A–H (steps, deliverables, stop conditions)
• source-index.md — src-NNN → talk learnings in ingest-into-skills

Optional: {SKILL_OUTPUT_DIR}/build-aiewf-2024-rag/

---

Step 0 — Pick workflow

Use the decision tree below. Open the matching section in workflows.md.

What is the user trying to do?
├─ Fix PDF/table RAG failures (layout ingest)           → A
├─ Add knowledge graphs (GraphRAG)                        → B
├─ Tune RAG with eval compass                             → C
├─ Harden LLM IO with Pydantic/Instructor                 → D
├─ Unified doc+vector on MongoDB                          → E
├─ Explore beyond RAG (EMT memory tokens)                 → F
├─ Vertical spec-heavy RAG (construction)                 → G
└─ Personal assistant structured data (minimize LLM)    → H

Stop summarizing once a workflow is identified — run its checklist.

---

Install

cp -r skills/build-aiewf-2024-rag ~/.claude/skills/
cp -r skills/build-aiewf-2024-rag ~/.cursor/skills/
cp -r skills/build-aiewf-2024-rag ~/.codex/skills/

From skills-i-use or ingest-into-skills (playlists/rag-llm-frameworks-aie-world-s-fair-2024/).

---

Cross-cutting rules

Rule	Source
Fix ingest before more retrieval tricks	[src-002 @ 1:21]
Eval-driven chunk/rerank changes	[src-004 @ 1:48]
Structured outputs via Pydantic	[src-005 @ 12:19]

Disputed steps: see source-index.md. Name workflow A–H; save artifacts to ./skill-outputs/build-aiewf-2024-rag/ when requested; do not auto-commit.

---

Invocation examples

@build-aiewf-2024-rag our RAG fails on tables in PDFs
when should we add GraphRAG?