Ai Training

Ai Technical Training

• Operator-first, not tool-first. We teach how to read signals, set minimal contracts, and run decisive escalations—independent of vendor hype cycles.

• Built for production. Templates for evaluations, runbooks, validators, and incident drills keep teams shipping reliable outcomes—not demos.

• Compounding improvement. Our Waste Monetization Framework (WMF) converts every “bad output” into tests, guardrails, and reusable fixes.

How the disciplines connect

• Thinking in AI — signal literacy. Learn to read systems, design lightweight contracts, and know when to escalate.

• Agile AI — cadence & loops. Work in thin slices, build evaluation harnesses, and use Definitions of Ready/Done for prompts and retrieval.

• Lean AI — waste → value. Apply WMF to convert failures into tests, guardrails, and knowledge assets that reduce rework.

• Ai CAGE Control — operations under turbulence. Operate safely with runbooks, validators, and incident drills that manage the 30%.

Course Catalog (Coming Soon)

Thinking in AI

Think like an operator, not a tool user. Map vendor specifics to shared abstractions: signals, contracts, checklists. Master the 70/30 mindset and practice when to invoke CAGE to keep outcomes on rails.

You’ll learn

• Signal literacy: distinguish model noise vs. actionable signals.

• Minimal contracts: precision prompts, retrieval constraints, data boundaries.

• Escalation triggers: how to detect drift and flip to human-in-the-loop.

Artifacts you take away

• Contract patterns (generation, retrieval, tool-use).

• Escalation checklist & decision tree.

• Signal log template for post-mortems and tuning.

Agile Ai

Apply Agile discipline to AI work: thin slices over big bets, short feedback loops, prompt/eval versioning, and release gates. Ship reliable 70% wins while containing the 30% through deliberate experiments.

You’ll learn

• Slicing AI features with measurable acceptance criteria.

• Evaluation harnesses (gold sets, heuristics, human review).

• DoR/DoD for prompts, retrieval, and tools; release gating with metrics.

Artifacts you take away

• Backlog templates for AI epics & spikes.

• Eval harness starter kit + tracking spreadsheet.

• Prompt/retrieval versioning scheme and change log.

Lean Ai

Use the Waste Monetization Framework (WMF) to turn “bad outputs” into assets. Capture → Classify → Convert → Codify → Compound. Reduce rework, control drift, and build a library of repeatable fixes.

You’ll learn

• Waste taxonomy for AI systems (defects, delays, rework, over-gen).

• Converting misses into tests, guardrails, and auto-fix playbooks.

• Cost-of-quality math to prioritize the highest-leverage fixes.

Artifacts you take away

• WMF board & labels for your tracker.

• Guardrail patterns (validators, refusal rules, safe tool-use).

• “Top 20 misses” library with fixes & tests.

Ai CAGE Control

Ground control for turbulent moments. Practice CAGE with validator agents, incident cards, and escalation checklists so teams can diagnose quickly, act safely, and return to stable flight.

You’ll learn

• Contracts: define scope, inputs, data rights, and acceptance criteria.

• Actions: safe tool invocation, rollback plans, and audit trails.

• Ground truth: sampling, adjudication, and drift monitoring.

• Escalation: incident drills, comms trees, and decision authorities.

Artifacts you take away

• Incident runbooks & tabletop drill pack.

• Validator agent blueprints (regex/ML/LLM mixed).

• CAGE “go/no-go” gate for releases.

Formats & Audience

• Formats: self-paced modules, cohort workshops, and on-site intensives.

• Audience: product leaders, architects, data/ML engineers, and operators responsible for real outcomes.

Outcomes you can expect

• Consistent 70% capture with evals and gating that prevent regressions.

• Controlled 30% execution under CAGE with faster incident recovery.

• A living library of tests, guardrails, and fixes that compounds in value.

• Repeatable playbooks that survive model updates and vendor changes.

Sample 2-Day Intensive

• Day 1 AM: Thinking in AI, contracts, and escalation drills.

• Day 1 PM: Agile AI slicing + build your eval harness.

• Day 2 AM: Lean AI WMF—convert your top 10 misses into tests/guards.

• Day 2 PM: CAGE incident simulation + release gate.