VOLUME 1 · THE THINKING LAYER
The AI Operator's Playbook
Models, Prompts & Grounding
From picking a model to shipping a grounded, tested workflow — without writing a line of code.
- 01Model Selection Matrix — reasoning vs instruct vs fast, cost-per-outcome, multi-model routing
- 02Context Windows & Token Economics — context rot, position bias, prompt caching, the Devanagari blowup
- 03Sampling Parameters — exact temperature values by task
- 04Prompt Architecture — rebuild one production prompt properly
- 05Reasoning Control — decomposing genuinely hard tasks
- 06Structured Output Engineering — extraction schemas that survive production
- 07Prompt Templating Systems — a real prompt registry, versioned
- 08Prompt Injection & Model Security — attack patterns, and a red-team pass on your own workflow
- 09Retrieval-Augmented Generation — measure and upgrade a RAG pipeline
- 10Long-Context vs RAG — compare three approaches on your corpus
- 11Memory Systems + Reference Vault
Actual pages from Volume 1
Arushi
Ashutosh Sehwag
Rupal Maloo
Sukwinder Singh Bassi
Yashwant Reddy
Priya Sharma
Shivkumar Mehta
Arushi Verma
Kiran Patel
Amita Sharma