Pablo Magariños

AI researcher working across industrial mathematics, aerospace engineering, world models, causal learning, and multi-agent intelligence.

My background in Aerospace Engineering provided a strong analytical and technical foundation, as well as a broad understanding of physics and the connections that link different scientific disciplines. During my degree, I developed a strong interest in artificial intelligence, motivated by longstanding questions about intelligence, the brain, and the nature of knowledge. That interest ultimately led me to Industrial Mathematics as a way to deepen my understanding of reality and contribute to AI from a more fundamental perspective.

Education

• MSc in Industrial Mathematics
  Universidad Carlos III de Madrid
• Degree in Aerospace Engineering
  Universidade de Vigo

Research Focus

My research is centred on the development of intelligence that genuinely understands its environment, rather than merely identifying loose statistical correlations in data.

The long-term objective of this work is twofold:

• To build AI systems capable of generating new knowledge and developing their own methods.
• To advance multi-agent AI toward a form of genuine swarm intelligence.

To pursue these goals, I work on topics such as:

• World models
• Causal learning
• Novel training paradigms

I also place strong value on building algorithms from first principles and testing them directly, including when experimentation produces unexpected or unsuccessful results.

Technical Profile

• Primary language: Python
• Current exploration: Julia, particularly for the different way it encourages mathematical and computational thinking
• Approach: Strong interest in understanding tools and libraries from the inside out, rather than using them as black boxes

Open Source

I am strongly committed to open source and open access. My projects are actively being developed in public and made openly available.