Skip to Content
From data to agent: how AI works and how to work with it

From data to agent: how AI works and how to work with it

An AI agent can hand you a well-organized report in eleven minutes, with figures and recommendations. The hard part is no longer producing it: it is knowing whether it is true. This course goes through, in eight blocks, what is inside an agent —data, probability, models, evaluation, language— so that you can read its work, check it and decide whether to sign off on it.

Private Course
Please sign in to contact responsible
Responsible Carlos Pascual
Last Update 2026-10-06
Completion Time 13 hours 15 minutes
Members 1
Basic Data, Analytics and AI
  • Block 1: The map: from asking the data to commissioning an agent
    3Lessons · 1 hr 10 min
    • Three questions to the data and what changes with AI
    • The team, the agent and the six-verb map
    • Block 1 exam
      10 xp
  • Block 2: Data: the raw material
    4Lessons · 1 hr 30 min
  • Block 3: Uncertainty: probability so you don't believe everything
    4Lessons · 1 hr 45 min
    • Probability: spreading belief across what can happen
    • Bayes and false alarms
    • Relationship, chance and cause
    • Block 3 exam
      10 xp
  • Block 4: Learning from data: what training means
    4Lessons · 1 hr 45 min
    • The smallest model: a line with two knobs
    • From the line to the neuron
    • The families of learning and the map of techniques
    • Block 4 exam
      10 xp
  • Block 5: Generalizing and evaluating: how to know if it works
    4Lessons · 1 hr 40 min
    • Memorizing is not learning
    • Hits, misses and what each one costs
    • Reading an evaluation without being fooled
    • Block 5 exam
      10 xp
  • Block 6: Language models: a text-prediction machine
    4Lessons · 1 hr 45 min
    • From text to vectors
    • How it generates: context and the next word
    • Why it is wrong with such confidence
    • Block 6 exam
      10 xp
  • Block 7: Agents: from answering to acting
    4Lessons · 1 hr 55 min
    • What an agent is: a model, tools and a loop
    • Giving it context and briefing it well
    • Reading what it has done and keeping a record
    • Block 7 exam
      10 xp
  • Block 8: Person and machine: dividing, supervising and answering for the work
    4Lessons · 1 hr 40 min
    • What gets delegated and what does not
    • Supervising without nodding off and without losing the craft
    • Answering for what you sign
    • Block 8 exam
      10 xp