Volume 3 — Model Lifecycle and Adaptation

How models are selected, modified, evaluated, compressed, deployed, and retired.

Reports

AI-ENG-G — Model Selection: Capability Fit, Deployment Fit & Failure Tolerance

Covers choosing models by task profile, reasoning depth, context length, tool use, modality, latency tolerance, cost ceiling, license constraints, language coverage, privacy requirements, hardware target, and acceptable failure modes. Teaches model choice as architecture, not leaderboard shopping.

AI-ENG-H — Model Adaptation: Fine-Tuning, LoRA, Preference Tuning & Distillation

Covers supervised fine-tuning, LoRA/QLoRA, preference tuning, domain adaptation, synthetic data generation, dataset design, adapter management, and distillation. Explains when adaptation improves behavior, when it merely overfits style, and when RAG or harness design would be cleaner.

AI-ENG-I — Regression Control: Model Registries, Rollouts & Behavioral Drift

Covers model registries, experiment tracking, artifact versioning, model cards, canary deploys, shadow testing, A/B tests, rollback procedures, and silent regression detection. Focuses on preserving production behavior as models, prompts, tools, corpora, and workflows evolve.

← Back to Canon Map