Why I’m Creating a Lab to Explore AI Behavior and Personality—And Where It’s Headed

AI has been on my mind for 30 years. From early research to consulting to AI art, I’m finally building the Savalera Agentic Lab to explore AI behavior, personality, and collaboration.

On February 17, we launched the Agentic Lab at our family business, Savalera.

This has been an idea on my mind for 30 years now. That’s a hack of a lot of time. Let me share a personal note on why I wanna do this and where the project is heading.

How It All Started

The idea first came to me in 1995, when I was studying computer science at university. In my third year, I took a subject called Artificial Intelligence.

To give some perspective, back then we studied Lisp as the language of AI—Python wasn’t mainstream yet. Neural networks were a separate subject, related but not fully integrated into AI.

We studied algorithms, game theory, problem-solving with graphs, NLP, semantic networks, and more. I was hooked immediately.

But from the start, I had my own idea. I’ve always been drawn to deep philosophical questions about existence, and since I first touched a computer at age 12, I knew I wanted to use computing to explore these questions someday.

I spent years thinking about:

  • What is consciousness? Is it innate, learned, or something in between?
  • What’s in a newborn’s mind? Are we blank slates, or do we come pre-configured?
  • Is the self a defense mechanism? Is it something that serves us, or does it hold us back?
  • Can we create consciousness artificially?
  • Can AI help us simulate human development and understand how we work?

These are hard questions, requiring a mix of philosophy, psychology, medicine, social sciences, spirituality, art and more.

That’s why I took up cognitive science in parallel with AI studies.

Along with two friends, I started my own research. For three years we worked on various topics and in our final thesis we created NALBISS (Natural Language-Based Intelligent Search System). My friends focused on search and NLP, while I worked on a structured semantic network—a system that encoded knowledge about the world in a way AI could process.

My work was greatly inspired by the Cyc project, which has since been discontinued in 2017.

The university was incredibly supportive. We had our own lab space, access to professors, and could continue our research throughout multiple semesters.

Our research was a bit controversial. We weren’t focused on pushing AI methodologies forward, we wanted to apply AI to real-world problems. We were enthusiastically building new stuff while not really advancing science. Besides the core work, I spent many hours talking to my cognitive science professor, asking high-level philosophical questions. I still remember him rolling his eyes sometimes.

After graduation, I felt like many of my core questions were left unanswered.

A PhD didn’t seem like the right path—universities were more focused on practical AI advancements, while I wanted to explore the more abstract aspect. Also, my research sponsors and professors didn’t really encouraged me to stay, so I decided to start my professional life instead.

From Consulting to AI Research

I joined Andersen Consulting (which later became Accenture) and have been working as a consultant ever since. I founded Savalera in 2013.

Helping organizations adopt new technologies and transform their businesses became my expertise. I enjoy consulting work a lot. Whatever I do in life I do it with a consultant’s mindset.

Teaching has also been a big part of my journey. I’ve taught Lisp at university, delivered corporate training courses (it’s still an active offering of Savalera), and even wrote a book on Docker and containerization in 2017.

After finishing a major transformation program in late 2022, I decided to invest more time in creative projects and return to the questions that kept me awake 30 years ago.

AI had made immense progress. The foundational work was done — I now have a once-in-a-lifetime opportunity to apply my ideas in ways that weren’t possible before.

I started working with AI models, explored AI art and NFTs, and built creative agents using fine-tuned open-source models to generate artworks about the human condition.

The theme of human experience, psychology, and self-awareness comes up a lot in my work. Over the years, I’ve seen burnout, stress, anxiety, and depression firsthand—both in my own life and in those around me. Managing large-scale transformation programs under high time pressure has shown me that understanding the human psyche is my most powerful skillset.

This realization deeply influences my art and my AI research.

The AI agents I built for art opened up new creative frontiers, and I realized I could bring both my professional expertise and artistic vision together by diving deeper into multi-agent AI research.

Why I Started the Agentic Lab

In 2025, I decided to formalize all my ambitions into the Agentic Lab at Savalera.

Here’s what I want to achieve:

  • Develop AI agents with real behavioral depth; agents that can adapt, evolve, and interact meaningfully.
  • Advance AI agent evaluation and benchmarking, focusing not just on performance but on behavior and personality.
  • Run experiments to understand human behavior using AI simulations to study decision-making, collaboration, and emotional intelligence.
  • Create art with deep human meaning; leveraging behavioral datasets from my research to drive artistic exploration.

My plan is to create practical results for people and businesses, I don’t necessarily expect to solve the big questions I asked as a student. But I do believe AI can help us get closer to understanding ourselves.

If this kind of research interests you, or if you want to collaborate, feel free to reach out.