I want to finish every course and tutorial in Anthropic's free learning platform, Claude Academy, in order. I'll write here about what I learn and build along the way.
Why now
After finishing my degree in software engineering, I worked as a frontend developer from 2019 to 2026, both full-time and freelance, for agencies and e-commerce companies. I built projects with many technologies, from traditional to modern ones.
While learning engineering practices and product development, I saw that communication is just as important a part of the work. So I turned my career toward communication and started working as a corporate communications specialist at a well-established law firm. There, I've gone beyond a traditional communications role and combined my technical background with my communication skills. Today I'm responsible for many things: managing the website and content management system, coordinating with agencies, tracking analytics and building AI-assisted content workflows.
Over the past year, AI has become part of my daily work. Sometimes I use it to create content from scratch, sometimes to research and analyze a topic. When I build a workflow, set up an automation or work on a presentation, it shows up at some point in the process. Claude, ChatGPT, Copilot, Legora and many other tools make my work easier and teach me new things at the same time.
What I watch for most when working with AI is hallucination: wrong information can sound just as fluent as the right answer. Then there's speed. When information comes this fast, it can make you a little blind, and it gets easier to accept an answer without questioning it. In fields like law and communication, where being accurate and consistent is essential, these two together mean every answer from AI needs a second look. Talking about slowing down may sound strange at a time when we're all trying to do everything faster. But I think that's exactly why, while we speed up, we also need to slow down and understand.
All of this taught me something: getting a good result once is easy. Getting it consistently, and knowing when to trust it, is a skill.
I saw this most clearly in a software project I've been building with Claude and Claude Code. In the project, I wear different hats, as if I were working with a software team. I treat engineering practices like system design, user experience, testing and documentation as separate roles and try to bring them into a structure. What excites me most is being able to apply these practices with AI, and to keep learning by finding and analyzing information quickly as I build.
When I'm curious about something, just using it isn't enough for me. I enjoy digging into the details to see how it works, then stepping back to the big picture and making it better. Now I want to really understand this skill, from how the models work to how applications are built on top of them, and to grow my engineering side in AI so I can keep up with this fast-moving field.
What I'm doing
Anthropic has a free learning platform called Claude Academy. It has courses and tutorials on AI fluency, Claude's apps, Claude Code, and building with the Claude API and the Model Context Protocol (MCP).
My goal is to finish all of it: 26 courses and 120 tutorials. Not in the order the site lists them, but in an order that moves from the fundamentals toward engineering.
The roadmap
- Foundations. How language models work, what they can and can't do, and the 4D framework for working with AI.
- The Claude app. Projects, artifacts, skills, connectors and research.
- Cowork and Claude Tag. Handing off longer tasks and working with AI as a team.
- Claude Code. Coding with an AI agent in the terminal: context files, skills, subagents, hooks and CI.
- Platform and API. Tool use, RAG, evaluations (evals), MCP and managed agents. This is the engineering core.
- Cloud providers. Using Claude through Amazon Bedrock and Google Cloud Vertex AI.
- Enterprise setup. How organizations roll out and manage AI tools.
- Industry-specific content. Education, nonprofits, finance and life sciences, for a wider view.
I'm most curious about phases 4 and 5. I already use Claude Code every day in the project I'm building. In these phases, I want to learn the methods behind what I've found through trial and error, and bring them back into the project.
How this series works
- A short LinkedIn post every week: one idea I found useful and how it worked when I tried it.
- A longer article here at the end of each phase: what I learned, what surprised me and what I'd do differently.
- Code on GitHub from phase 4 onward, when the courses move into building. Each project will be my own, with a short write-up.
I'm sure there are many people who use these tools, and many who know them much better than I do. What I want to do here is understand the tools I use every day from the ground up, and write about them through the eyes of both an engineer and a communicator. So I won't summarize the lessons; they're free and open to everyone. Instead, I'll write about what happens when I put what I learn into practice.
What I took from the first lesson
The first lesson of the series, Introduction to AI Fluency, defines AI fluency as the ability to work with AI in ways that are effective, efficient, ethical and safe, and groups it into four areas: Delegation, Description, Discernment and Diligence. The part that stayed with me most was the idea of changing how we think about AI so we can learn to think with AI.
While watching, I realized I had already been applying these four steps without noticing. Writing this article was the same: I got help from AI for the draft, explained what I wanted, reviewed the text with my own eyes and changed it many times, and in the end chose to say so openly. The lesson gave a name and a framework to something I was already doing. That's exactly what I want from this series: to understand what I've been doing by intuition.
Follow along
If you're learning AI too, or thinking about starting, follow the series and tell me where you are. Let's compare notes. If this series is useful to anyone coming to AI from a different field like me, or to someone just starting out, that would make me happy.
You can find the academy at academy.claude.com.
I wrote this article myself and used Claude for editing. The ideas, experiences and mistakes are my own.
From Using AI to Understanding It · Part 0 · Next: Part 1, Foundations
