Building reliable Agentic AI systems
Agentic AI moves beyond prompt-to-text. This article shows how LangGraph orchestrates stateful, feedback-driven workflows to turn news retrieval into a fact-checked, LinkedIn-ready post, with memory, conditional logic, and verification.
AI and the price of FOMO: Innovation without strategy
Fueled by the fear of missing out, many companies have jumped head-first into AI. The pattern is familiar: rushing into deployments, inflated expectations, hurried investments, and returns that fall short of the hype. In this post, we take a step back and look at what history can teach us about making AI work — how to avoid the pitfalls and maximise the chances of real success when bringing AI into business.
Exploring the boundaries of AI
In this post we explore the limits of AI development from two perspectives: a popular-science angle, reflecting on the deterministic nature of algorithms and the notion of intelligence, and a more technical approach, examining the mathematical and computational boundaries that define how far AI can go.
Bayes’ theorem and its role in AI
Bayes’ Theorem is more than a formula: it’s a way of updating what we believe when new evidence appears. From spam filters to medical diagnostics, it underpins many AI methods. In this post, we’ll see how machine learning uses Bayes to deal with uncertainty with both rigour and efficiency.
Layers of power: Inside neural networks
In this article, we unravel the fascinating process by which machines adjust their parameters to recognise handwritten digits. With visual examples, accessible mathematical explanations, and Python code snippets, we will show you the ‘behind the scenes’ of one of the most common applications of artificial intelligence.
Artificial Neurons and Perceptrons: The Foundation of Neural Networks
Translation coming soon. In this article, we unravel the fascinating process by which machines adjust their parameters to recognise handwritten digits. With visual examples, accessible mathematical explanations, and Python code snippets, we will show you the ‘behind the scenes’ of one of the most common applications of artificial intelligence.
Europe’s AI dilemma – innovate or regulate?
The European Union has implemented the Artificial Intelligence Act, a regulation designed to govern the use of AI without stifling innovation. How will this legislation affect Europe’s competitiveness? In this article, we examine the balance between regulation and technological progress.
Anatomy of an AI’s algorithm
Translation coming soon. In this entry, we shall explore, at different levels of difficulty, how ML algorithms adjust their parameters to solve problems. We shall see that, at its core, it all comes down to an error minimisation problem.
The importance of language in AI
Let us explore the dilemma and possible consequences of using anthropomorphic expressions in the context of AI.
DeepSeek-R1: Another tall tale from China?
In this entry, we explore the key differences between the new DeepSeek-R1 model and the well-established OpenAI-o1, examining its architecture, energy consumption, and overall reliability.