Copyright (c) 2026 MindMesh Academy. All rights reserved. This content is proprietary and may not be reproduced or distributed without permission.
7.2. Quick Reference
The Master Decision Filter
Run this on every scenario before reading the options closely.
Deployment Type Quick-Reference
The Lever Selection Table (highest-value distinction)
| Need | Lever | Never use |
|---|---|---|
| Steer one response (tone, format, this turn) | Prompting | — |
| Supply private/fresh facts | Grounding (RAG) | Fine-tuning (won't add facts) |
| Consistent style/format across calls | Fine-tuning | Grounding (it's not facts) |
| Take an action / call a system | Function/connected tool | Code Interpreter (sandbox only) |
| Analyze data, run math, make charts | Code Interpreter | Function tool for built-in analysis |
Control-to-Threat Quick Map
| Threat / need | Control |
|---|---|
| Runaway token spend (prevent) | TPM ceiling on deployment |
| Spend awareness (notify) | Budget alert |
| Jailbreak / prompt injection | Prompt Shields |
| Harmful content (hate/sexual/violence/self-harm) | Harm-category filter + severity |
| Model invents facts not in source | Groundedness detection |
| Sensitive data before an LLM | Language PII redaction (preprocess) |
| Authenticate without secrets | Managed identity (DefaultAzureCredential) |
| Over-permissioned runtime identity | Least-privilege RBAC (OpenAI User) |
| Control/audit encryption keys | Customer-managed keys (CMK/BYOK) |
Search Strategy Quick-Reference
| Query trait | Use |
|---|---|
| Exact codes/IDs + concepts | Hybrid (keyword + vector) |
| Concept/paraphrase only | Vector |
| Need top-result precision/nuance | Add semantic ranking (cross-encoder) |
| Keep index fresh, no custom code | Indexer + integrated vectorization |
Written byAlvin Varughese
Founder•18 professional certifications