In this episode, Alexandra Spalato shares how she moved from software engineering into entrepreneurship, focusing on automation, AI, and community-driven tools like n8n. She talks about managing AI expectations with clients, building products around real needs, and why community involvement is crucial when working with open-source technologies.We also discuss her journey on YouTube, the challenge of creating content that balances depth with accessibility, and why click-driven algorithms make it harder to stay authentic. Alexandra closes by reflecting on her future plans, including small SaaS products and more meaningful, human-centered digital experiences.

In this episode, Rosaria Silipo and Alessandro Romano break down what makes AI agents different from simple prompts and why agentic AI is becoming essential for real automation. We discuss multi-agent systems, orchestration, decision-making, and where agents make a real impact—from dynamic tasks to practical applications such as monitoring, research, and execution.You’ll learn when it makes sense to use agents, the risks involved (compliance, transparency, control), and what future standards may look like.

Follow us for more content: https://mydataguest.substack.com/

What’s the difference between training an LLM from scratch and fine-tuning one that already exists?
In this episode, we unpack the full story: the data, the GPUs, the cost, the workflows, and the practical reasons why most companies choose fine-tuning over full training. We also explain how fine-tuned models can adapt to new styles, languages, and business needs.

A simple and direct conversation about how modern AI models are actually built and shaped.