Impulsa IA · Academy
Master AI, one task at a time.
A structured path of 6 levels and 172 hands-on tasks to take you from prompt basics to designing autonomous AI agents with Claude.
Level 1 includes 20 available written lessons, exercises and quizzes. Videos and PDF handouts are still in preparation.
Every certificate is verifiable:JCA-XXXX-XXXX-XXXXLevel 1 preview
AI fundamentals and prompting
Understanding generative AI
- 1N1-T01What are generative AI and large language models?30′
- 2N1-T02How a language model generates text: tokens and the context window30′
- 3N1-T03Capabilities and limits: hallucinations, knowledge cutoff, bias35′
- 4N1-T04Overview of AI assistants and introduction to Claude30′
- 5N1-T05Using AI responsibly: privacy and ethics35′
- 6
- levels
- 172
- hands-on tasks
- 132
- estimated hours of practice
- 3
- languages
Practical skills
Learn today. Put it to work tomorrow.
From your first prompts to advanced projects, build tools that help in your everyday work.
Write with clarity
Prepare messages and documents tailored to your work.
Structure information
Turn text into summaries, tables and usable data.
Build your own tools
Progress toward assistants and automations tailored to your needs.
Six progressive levels
A path in six levels
From your first prompt to autonomous agents. Every level ends with a project reviewed by an instructor.
- 01
Level 1
AI fundamentals and prompting
Understand how generative AI works and master prompting techniques to get reliable, useful answers from Claude.
20 tasks3 projects≈ 15 hAvailableExplore this level - 02
Level 2
Claude Code and the command line
Install and master Claude Code in the terminal: sessions, CLAUDE.md memory, configuration, permissions, plan mode, git and non-interactive mode.
32 tasks3 projects≈ 22 hIn preparationExplore this level - 03
Level 3
Tools and integrations
Connect Claude to your ecosystem: MCP servers, the Claude API, GitHub, databases and automation.
35 tasks3 projects≈ 27 hIn preparationExplore this level - 04
Level 4
Subagents and multi-agent orchestration
Design specialized subagents, orchestrate multiple agents and build custom agents with the Claude Agent SDK.
25 tasks3 projects≈ 20 hIn preparationExplore this level - 05
Level 5
Skills, hooks, permissions and security
Extend Claude Code with skills, hooks and plugins, and secure its use: permissions, sandboxing, prompt injection, secrets and review.
35 tasks3 projects≈ 25 hIn preparationExplore this level - 06
Level 6
Autonomy: long-running agents and deployment
Entrust entire projects to autonomous agents: planning, long-running execution, evaluation, deployment and a capstone project.
25 tasks3 projects≈ 23 hIn preparationExplore this level
Anatomy of a task
Every task follows the same structure, designed for learning by doing.
- 01Objectives
- 02Lesson
- 03Examples
- 04Demonstration
- 05Exercise
- 06Deliverable
- 07Quiz
- 08Success criteria
Learn by doing
Each task combines a lesson, a demonstration, an exercise and a concrete deliverable.
Measured progress
Quizzes graded on the server, projects reviewed by an instructor, and an 80% passing threshold.
Verifiable certificate
Every level you complete earns you an internal certificate that can be verified online with a QR code.
Verifiable certificate
A certificate anyone can verify
Each completed level earns an internal certificate with a unique ID and a QR code. Anyone can confirm its authenticity online.
- Unique ID in the JCA-XXXX-XXXX-XXXX format
- QR code linking to the public verification page
- Minimal data shown: name, level, date, status
Certificate of Achievement
certifies that
Your name
has successfully completed
Level 1 — AI fundamentals and prompting
JCA-XXXX-XXXX-XXXX