Generative AI Engineering
Signature Course No. 216 — A Signature AI Socratic Coaching Course.
A New Kind of Learning
Socratic coaching, delivered by an agentic AI coach.
Not a course you watch. An AI coach you work with.
The Lighthouse Coach begins by discovering how you think about the subject, teaches what is missing, and gives you realistic practice before asking you to prove that you can use it. You may test out of concepts you already understand, while capstones turn the learning into a practical plan, decision, conversation, or other real-world result.
The curriculum is deliberately authored; the AI delivers it through coaching. It is not an AI randomly generating lessons.
A coach that teaches by asking — not by lecturing
The Lighthouse Coach is an AI Socratic coach built on Cognosco Learning's Generative AI Engineering course, in the AI at Work category. It does not play videos at you. It holds a conversation: it asks you a question, listens to your answer, names the misconception it hears, affirms what you got right, and asks the next question — the way a great coach works in the room.
Anyone can get fluent text back from a generative model. That is not engineering. Engineering begins where the demo ends: when the output has to be right often enough to stake a customer's claim, a patient's instruction sheet, or a company's money on it — and when the bill and the wait are part of the specification.
You learn the mechanism first, then the design surface. A generative system is a sampler, conditioned by everything you feed it, whose behaviour is a distribution and not a function. You will write prompts the way engineers write specifications, build context the way engineers build inputs, force output into shapes other software can trust, and set decoding controls like any control with consequences.
Then the half demos never show. Evaluating output you cannot fully predict, judging it at scale without fooling yourself, releasing changes without silently breaking what worked, guarding the perimeter against honest traffic and adversaries alike, budgeting cost and latency before the invoice and the timeout teach you, and watching a running system so its failures find you first.
It works in two movements. First, Module 1 — Designing Generation — work through Concepts 1–9, finishing with The Design Pack; then, Module 2 — Proving, Protecting, and Paying for Generation — work through Concepts 10–18, finishing with The Release Dossier.
Both capstones produce real artifacts — The Design Pack, a complete, gradable design for Sycamore & Strand's rebuilt claims-summary generator; and The Release Dossier, which takes the support-draft assistant from candidate to a signed shipping decision — evidence, gate, perimeter, budgets computed line by line, and the decision in its five parts.
The honesty discipline is part of the course. No commercial model's capabilities, prices, limits, or rankings are named as facts anywhere; the teaching is at the level of capability and mechanism, with a direction to check current provider documentation wherever a product fact would matter. Every worked figure is an invented scenario figure, labelled as such, and every performance standard is stated as a rate over a named set — never as a single successful run.
The course is self-paced and always on — roughly 30 minutes per concept, about 9 hours in full, taken in as many sittings as you like. You can type your answers, or talk out loud: the Voice Coach speaks its questions and listens to your spoken replies. Learners who already know the material can test out concept by concept.
Generative AI Engineering is part of The Lighthouse Signature Series — the online counterpart to the one-day, in-person learning Cognosco Learning has delivered since 2002.
Course Facts
One full concept, coached end to end — free: verify your email and we’ll send you a link to start.
Already enrolled? Enter the course →
Diagnose → Learn & Practice → Prove → Capstone
Every learner moves through the same discipline a Cognosco instructor would insist on in the room.
Diagnose
The Coach starts where the course starts: with what you actually bring. A generative system does not have an output for an input — it has a distribution of outputs, and the first questions are always how often, on what inputs, judged how. Every concept opens with three questions that locate where you actually stand, and your answers shape how deeply the Coach works the concept with you. Everyone begins with Module 1 — no skipping the foundations you only think you have.
Learn & Practice
Work each concept in conversation with the Coach — spoken or written — then practice it against realistic situations until it holds up.
Prove
Every concept ends in proof: answer the check cold and demonstrate the skill. Know it already? Test out and take credit immediately.
Capstone
Finish with two capstones — The Design Pack for Module 1 and The Release Dossier for Module 2 — that pull the whole course together, and earn the certificate.
What You'll Learn
A distribution, not a function. How a generative system actually produces output — conditioning and sampling — and why one impressive demo is not evidence about the system you have.
Module 1 designs generation. The model bounded as a component, the prompt as a specification, context engineered to a budget, output forced into validated structure, decoding set per job, grounding at its seam, pipelines decomposed only where cutting pays, and the failure modes named in a register.
Module 2 proves it before it ships. Reference sets and rubrics, judges calibrated against human judgement before they are trusted, releases gated as events, and guardrails placed where holding is mechanical — against honest traffic and adversaries alike.
And it pays for it, watches it, and signs it. Cost from the full act and latency along the critical path, observability that reconstructs any disputed output, and a shipping decision in five parts — claim, evidence, bounds, watch, signature.
What you will learn
- Explain how a generative system produces output — conditioning, sampling, and why its behaviour is a distribution, not a function — and choose and bound a foundation model on capability evidence, not reputation
- Write prompts as specifications and engineer context — what goes in, in what order, and what stays out — and force generation into structured, validated output that downstream software can trust
- Set decoding controls per job, ground a generator in retrieved material as one component among others, and build multi-step pipelines that name the intrinsic failure modes of generation before they ship
- Evaluate output with test sets, rubrics, and judges you have yourself checked — and ship, guard, budget, and monitor the result under a signed standard
Everything in the course
- All 18 concepts of Generative AI Engineering, in 2 modules
- A starting-point check first — the Coach finds where you actually stand before teaching
- Learn, Practice, Prove for every concept — no passive video watching
- A test-out on every concept: prove you know it and move on
- The Voice Coach: talk with the Coach out loud — it speaks, listens, and replies
- A personal Learner Record of your work across the course
- Two capstone exercises: The Design Pack and The Release Dossier
- A printable session summary whenever you pause
- The Lighthouse Coach completion certificate (Signature Series)
- Your progress is saved as you go — stop anytime, resume where you left off
Built for people who ship generative systems
Software engineers — handed generative features to build, who need the specialist discipline beneath the API call: design, evaluation, guardrails, budgets, operations.
Machine-learning and data practitioners — who know models from the fitting side and need the production side: how a fitted model becomes a dependable component in a system others rely on.
Technical leads and architects, and builders of internal tools — who choose components, set budgets, review designs, and decide what evidence is enough before a feature ships, and who generate real work product with these systems in operations, support, legal, finance, and service teams. The course assumes you can read code, call an API, and reason about a system in parts; it assumes no prior specialty in generative systems.
Prefer to learn live and in person? The same material anchors Cognosco's one-day programs — browse the in-person catalog.
Start with The Design Pack
$195 — introductory price. One payment. No subscription, no password to manage.
Here is exactly what happens when you enroll:
- Pay securely with Stripe — you are taken straight into the course
- Create your learner profile and begin the Module 1 starting-point check
- Save the access key you are given — it restores your enrollment on any device
- Learn at your pace; your progress is stored in your browser on your device
For the full voice experience, use Chrome and allow the microphone when asked. Typing your answers works everywhere, all the time. Questions before you enroll? Email hello@cognoscolearning.com.
More Signature Courses are coming
Generative AI Engineering is Signature Course No. 216. Leave your email and we'll send you news of future Signature Courses and Lighthouse updates as they are announced — nothing else.