Online course on AI and data
From data to agent
How AI works and how to work with it. Eight blocks to understand what is inside an agent, from data to language, and to learn how to brief it, read what it does and answer for the result.
View the course See the syllabus
- 8 blocks and 23 lessons, online and self-paced, in Spanish
- Beginner level: no coding or math background needed
- 12 interactive exercises and the case data to download
- An exam with explained answers at the end of every block
- It opens both learning paths and stands on its own
About the course
A report in eleven minutes. Would you sign it?
The hard part is no longer doing the analysis. It is knowing whether it is true.
An AI agent can hand you a tidy report, with figures and recommendations, in eleven minutes. This course walks through what is inside an agent (data, probability, models, evaluation and language) so you can read its work, check it and decide whether to sign it.
Everything revolves around a single case: the report an agent wrote for the management of a bakery chain, made up for the course, about falling sales. Each block teaches you to read one layer of that report. The data is yours: you download it and check every figure with your own AI assistant.
Behind every idea in today's AI there is a classic concept that explains it: the neuron is a logistic regression, generating text is sampling from a distribution, evaluating a system means keeping back an exam it has not seen. Here you learn both, the concept and the bridge, with small numbers and no blackboard formulas.
This is not a course about one tool, or about prompt-writing tricks.
Who it's for
Who is this course for?
Working professionals who decide with data and are about to hand work to AI agents.
You decide with data
You work in business, finance, operations, marketing or consulting, and you are about to hand work to AI agents.
You use AI every day
You are an analyst or have a technical background and want to understand why it works, when it fails and how it is evaluated.
You are starting a learning path
It is the first course of both Learning Paths in Data, Analytics and AI: it covers the whole route before you go deep.
Learning outcomes
By the end you will be able to
Explain how it works
Explain what data, a distribution, a model, an evaluation, a language model and an agent are, and how each one builds on the one before.
Name what lies behind
Name the classic concept behind ten ideas in today's AI, and say what it explains and what it does not.
Brief and verify
Hand a small analysis to an assistant with a five-part brief and verify it with a four-check record.
Read a trace
Read an agent's trace and find the step where it goes wrong.
Decide what to delegate
Decide which part of a job can be delegated to an agent, with how much review, and who answers for it.
Syllabus
The syllabus, block by block
Eight blocks and 23 lessons: from data to agent, with a single case read layer by layer.
The map: from querying data to briefing an agent
The questions you ask of data, what has changed with pretrained models and agents, the roles on a data team and how you study here. It introduces the case that runs through the whole course: the report an agent wrote about falling sales.
Data: the raw material
What shape data takes, how to summarize it without fooling yourself, and where data comes from and who it belongs to. First reading of the agent's report: which table it used and what summaries it gave.
Uncertainty: probability, so you don't believe everything
Spreading belief across what might happen, flipping a probability with Bayes, and telling a real relationship from one manufactured by chance or by a third variable. Second reading of the report: what it claims and what the data lets you claim.
Learning from data: what training means
A model is a formula with dials; training means turning them to be wrong less often. From the straight line to the neuron, the three ways of learning and the map of techniques. Third reading of the report: the model the agent fitted.
Generalizing and evaluating: how to know whether it works
Memorizing is not learning; you don't show the exam beforehand; there are two ways to be wrong and they don't cost the same. Fourth reading of the report: how the agent evaluated its own model.
Language models: a machine for predicting text
From text to vectors, from vectors to the next word, and from there to why a model can be so confidently wrong. Fifth reading of the report: how the text around the figures was written.
Agents: from answering to acting
What turns a model into an agent, what it needs to know about your problem and how to read what it has done. Sixth reading of the report: the trace, step by step, and the five errors it was hiding.
Person and machine: sharing out the work, supervising and answering for it
What you delegate and what you don't, how to supervise without dozing off or losing your craft, and who answers for what gets signed. Final reading of the report: would you sign it?
Learning by doing
Try it: an exercise from the course
Four tables with the same summary statistics. Are they the same? Place your bet, plot them and drag a point to watch the mean, the correlation and the fitted line change. The exercise is in Spanish: try it right here, no sign-up needed.
Methodology
How you learn
Online, self-paced
You take it in The Singular Academy's Virtual Classroom, in Spanish and with no cohort: you set the pace.
One case, read eight times
Each lesson opens with a part of the agent's report and gives you the ideas to judge it.
You turn the dials
Twelve interactive exercises: you fit a line, work out a neuron by hand or follow an agent's loop, step by step.
You practice with your AI assistant
Every block includes an exercise with your AI assistant on the case data; any assistant will do. What you hand in is not what the assistant did, but your reading and your verification.
Explained exam answers
Every block ends with an exam. Each answer, right or wrong, comes with the reason why.
No coding required
You don't need to code or have a math background: you get into the subject through examples and small numbers, not proofs.
Why The Singular Academy?
Built on a business case
You learn by reading one case from start to finish, not from scattered examples.
From concept to bridge
First the classic concept, then the AI idea it unlocks, always on the same case.
No dates or timetables
Log in to the Virtual Classroom whenever you like and move ahead one short unit at a time.
Backed by Singular Beacon
A consultancy specializing in digitalization and data strategy.
Key facts
Practical information
- FormatOnline, self-paced
- AccessBy invitation
- LanguageSpanish
- LevelBeginner
- PrerequisitesNo coding required
- Content8 blocks · 23 lessons · 12 interactive exercises · downloadable case data · an exam with explained answers per block
- WhereThe Singular Academy Virtual Classroom
Access
Request your access
The course is listed in the catalog and its welcome lesson is open to everyone. The rest unlocks when The Singular Academy grants you access: request it from the Virtual Classroom or through the contact form.
Course access
Access by invitation
Request access- Eight blocks and 23 lessons in the Virtual Classroom
- Twelve interactive exercises where you turn the dials
- The case data, to check every figure with your assistant
- An exam with explained answers in every block
The learning paths
The first course of both learning paths
"From Data to Agent" opens The Singular Academy's two nine-month learning paths: Artificial Intelligence for Professionals and Data Analytics with an AI Copilot.
The whole route, in one go
In eight blocks, this course covers the whole route from data to agent. The courses in both learning paths pick it up and go deeper, from November to June.
It stands on its own
You can take it without continuing along either learning path. And if you do continue, you have already completed the opening course of both.
Learn to stand behind what you sign
Start with the welcome lesson, which is open to everyone, and request access to the rest of the course.