Valencia, Spain
© 2026 Germán Mallo
Germán Mallo Data Science · Artificial Intelligence · Development
+
PERSONAL BLOG
Experiments, projects and learning
Are you interested in AI?
Let's talk

I'm Germán Mallo, a graduate in Data Science from the Polytechnic University of Valencia. I'm passionate about artificial intelligence and software development, and I enjoy creating tools that solve real-world problems, even small ones.
This is my personal space: here I share what I'm learning, the experiments I conduct with AI models, APIs, and automation, and the projects I've worked on. I write about what interests me and what I struggle to understand, because explaining it is the best way I know to learn it.
The last thing I wrote
Projects I have worked on
TECHNICAL PORTFOLIO
Automation of food contracts for IBEROCEBOLLAS Automation and Development
Sentinel Enervia Automation and Development
Salvest Gestiones | Invoice automation with AI Artificial intelligence
PostPilot AI: Data-driven content generation Automation and Development
EnerviApp AI Assistant Web Design
My learning process is the same every time: understand how something works, try it out with my own hands, and share what I've discovered.
01 — Learn
I read documentation, papers, and articles to understand how things work on the inside.
02 — Experiment
I test models, APIs, and tools by building small prototypes.
03 — Share
I write about what I learn to organize my thoughts and in case it's helpful to someone else.
CURIOSITY
I like to know what's behind each tool: how a language model works, why a RAG system fails, or what a new library actually brings to the table. Before using anything, I try to understand it.
LABORATORY
The best way I know to learn is by building. I set up proofs of concept, compare models, automate my daily tasks, and make enough mistakes to understand what works and what doesn't.
BLOG
Documenting what I do forces me to explain it well. On the blog, I share experiments, lessons learned, mistakes, and reflections on technology, from the perspective of someone who continues to learn every day.
LEARN, TEST AND SHARE: THIS IS HOW EACH ARTICLE IS BORN
What interests me most
Artificial intelligence
Language models, agents, RAG systems, and generative AI. I'm interested in understanding how they work internally, comparing them, and seeing how far they can go when applied to specific cases.
Data Science
My background includes: exploratory analysis, predictive models, time series, visualization, and everything involved in turning data into reliable conclusions.
Automation and Development
Python, APIs, scripts, and small applications. I enjoy automating my daily tasks and building tools that I actually end up using.
Tools and technologies that I use
The tools, languages, and platforms I use most for programming, data analysis, and experimenting with artificial intelligence.
ChatGPT
Generative AI and development
Claude
Generative AI and analytics
Gemini
Multimodal AI and development
GitHub
Version control
Notion
Organization and documentation
Cloudflare
Web security and performance
Python
Development, data and AI
Docker
Containers and deployment
PHP
Backend web development
The questions I receive most often about AI
- Intelligent Assistants
- Automations
- Data Science
- Artificial intelligence
- My opinion
Where do I start if I want to learn artificial intelligence?
My recommendation: learn Python and the basics of statistics, experiment with the APIs of current models, and build something small that you find useful. Understanding how embeddings, transformers, and techniques like RAG work comes more quickly and effectively when you apply them to your own project.
What is an AI agent and how does it differ from a chatbot?
A chatbot answers your questions. An agent, however, can decide what steps to take and use tools (search the internet, run code, consult an API, edit files, etc.) to complete a task on its own. It's one of the topics that interests me most and with which I experiment the most.
Is it better to use AI models locally or in the cloud?
For most uses, the cloud wins in terms of quality and convenience. Local models make sense if privacy is paramount, you want to experiment without usage costs, or work offline, but you need a computer with a good GPU and are willing to accept somewhat less powerful models. I explain this in more detail in one of my blog posts.
Can I trust the answers I get from an AI?
Proceed with caution. Language models can certainly fabricate data (the infamous hallucinations). For important matters, ask for sources, verify the information, and if it's code, test it. It's a magnificent tool for thinking and moving forward faster, not an oracle.
Will AI replace programmers?
I don't think so, but it is changing the way we program a lot. AI is writing code better and better; what still makes the difference is understanding the problem, designing a good solution, and knowing how to review what it generates. The programmer who learns to work with it has an advantage over the one who ignores it.
Is it worth paying for AI?
If you use it daily for studying, programming, or work, then yes, paid versions usually do: they give you access to the best models, higher usage limits, and features like file analysis. If you only use it occasionally, the free versions are more than enough. Before paying for multiple subscriptions, choose one and make the most of it.