Completion certificate
Your effort, on record.
A preview of your course milestone. Completion reflects the episodes you mark complete; it is not an externally accredited qualification.
For software developers building with AI
Focused lessons that respect your time. See how the ideas connect, then apply them in your own projects.
Shaped by experience at
$347 one-time · Lifetime accessOffer valid through October 31, 2026 (US Pacific Time). $647 from November 1.
All 28 courses. Every course’s practice skill.
Fresh technical content every month.
The video could not play. Open it directly or explore a free episode below.
Start learning. No sign-up.
Free mini-courseComplete course · 16:47
Follow a fictional scheduling assistant’s double booking. Catch the failure with code checks, a clear rubric and an eval suite.
The whole course. No email. No account.
The video couldn’t load. Open the free video directly.
Follow the failure. Build the check.16:47 · English captions
A confident answer can hide a broken schedule.
Code for hard rules. Rubrics for judgment.
Catch regressions before the next release.
Inside the mini-course
Catch a confident claim with saved-state checks, repair the schedule, and rerun every case. Take the mini-course from watching to doing.
YOUR FIRST MISSION
Find a double booking and a false coverage report in the offline practice cases. Repair the saved schedule, then check what improved and what still needs human review.
You’ll make: FIRST_WIN.md · EVAL_REPORT.md · eval_results.json · RUBRIC.md
Included with this free mini-course. No account, API key or paid tool required. Runs offline with Python 3.9+.
~/.codex/skills/ or ~/.claude/skills/.vivid-ai-evals-mini-course for your first eval.You can also run the helper directly from its folder:
python3 scripts/vivid.py firstwin /path/to/projectOriginal fictional cases, deterministic checks and a human-review rubric. The tool does not call a model or execute your project code. It writes practice reports inside your chosen project’s .vivid/vivid-ai-evals-mini-course/ folder. Prepared replays teach the method; they do not establish production reliability. Files you ask an AI agent to read may go to its model provider.
A claim is not proof. A fluent message can say that every station is covered while the saved schedule contains overlapping bookings. Evaluate the state and the trace, not just the assistant’s wording.
Match the grader to the question. Code can check availability, qualifications and consent rules. A clear rubric helps people and model judges assess whether a message explains a gap and offers a safe next action.
Rerun after every repair. Check the cases that improved and the cases that regressed. Separate development cases from held-back checks, and report counts and failure types.
The film uses prepared teaching examples and real code checks. Its model-judge verdicts and reviewer labels are illustrations, not live model measurements. This complete, standalone free mini-course is separate from the paid course catalog.
The video couldn’t load. Open the free video directly, or use the written lesson below.
Sample episode · AI Agent Architecture
Your model doesn’t know your docs. See where retrieval fits between a user’s question and the model call.
A practical look at context limits, relevant evidence and retrieval.
Training knowledge is not your project’s knowledge. A model’s general training does not include your private repository or internal policies. To answer a question about those documents, your application must supply the relevant information when it makes the model call.
Context is a budget. Instructions, conversation, retrieved material and the response all need room. Token limits vary by model; check the model you actually use. Sending the entire document collection is rarely a useful first design.
Relevance matters as much as capacity. Unrelated facts make the useful evidence harder to isolate. Start by selecting documents that address the question, preserve their source IDs, and define what to do when there is no evidence.
Retrieval is a step, not the answer. Question → retrieve relevant snippets → build context → ask the model → check the answer against its sources.
From a first win to your next build
Build with 28 focused courses in agents, evaluation, retrieval and more. Your one-time purchase includes the practice skill for each course you own; the full Vault unlocks every paid-course skill. Each skill gives you a concrete first task for your own project. Courses and technical content are updated every month.
The Concept Galaxy
See how each lesson fits into the bigger picture. Scroll between concepts, follow their connections to related episodes, and see your progress as you learn.
Make the whole learning path yours.
$347 one-time · All 28 courses + practice skillsOffer valid through October 31, 2026 (US Pacific Time). $647 from November 1.
Explore our AI courses, discover new topics, and continue learning at your own pace.
We don't have student quotes, success counts, or revenue figures to show you — so we show you the courses instead. Start exploring AI topics with our open learning resources, then discover more courses in Explore. More first episodes are free with a login. Verify the quality yourself before paying anything. That's the whole trust model, plus a named seller and a 14-day money-back guarantee.
Your milestones, made visible
Preview a course completion certificate, a milestone badge, and a purchase receipt.
Completion certificate
A preview of your course milestone. Completion reflects the episodes you mark complete; it is not an externally accredited qualification.
Milestone badge
A preview of the course completion badge: a small reminder of the concepts you have connected and the progress you have made.
Receipt & reimbursement
Planning to use a learning allowance? Check your employer’s policy first. Paddle emails your actual purchase receipt; prices and tax depend on your checkout.
These are samples. They contain fictional learner details. The receipt preview is not a tax invoice and cannot support an expense claim. Use the real document issued by Paddle after payment; reimbursement is subject to your employer’s approval.
The full AI Vault
Go from understanding the ideas to building with them. Take the whole collection, connect the concepts, and keep putting what you learn to work.
$347USD · one-time paymentOffer valid through October 31, 2026 (US Pacific Time). $647 from November 1.
Lifetime access. No subscription.
14-day money-back guarantee.
Secure checkout by Paddle. No account needed to buy. Any applicable tax is shown before payment.
See all purchase options →