Guardrails for generative AI: safety barriers in production
Guardrails for generative AI: what they are, input, output and action types, defence against prompt injection, common mistakes and how to validate them.
AI Forge
AI Forge | Think. Build. Evolve. Forge the Future | Artificial intelligence
Welcome to AI Forge, a space designed to dig into the engineering and practical application of Artificial Intelligence. In this corner of my blog, we move away from surface-level theory to focus on building real solutions, the technical deployment of advanced models and hands-on experimentation with code.
Through this digital lab, I explore how the engineering and practical application of Artificial Intelligence can turn complex workflows into efficient tools. Here we forge the future by analysing architectures, testing concepts and sharing first-hand experiences in developing intelligent software and autonomous systems.
If your organisation is integrating AI into real projects and you need to orchestrate the deployment, I invite you to explore how to hire an AI Project Manager or learn about my background.
-
-
Conformity assessment for high-risk AI systems
Conformity assessment for high-risk AI systems: what it is, the requirements assessed, internal vs notified-body routes and how to prepare from the design stage.
-
Context engineering: the successor to prompt engineering in AI projects
What context engineering is, why it surpasses prompt engineering and how to design the context a model receives to gain reliability and control.
-
GPAI obligations under the EU AI Act: what changes in August 2026
A review of the EU AI Act obligations that come into force in August 2026, who they affect and what to do so you don’t fall behind on compliance.
-
ISO 42001: how to implement an AI management system in the enterprise
What ISO 42001 is, how to implement an AI management system step by step and how it relates to the EU AI Act in regulated environments.
-
EU register of AI systems (Article 71): beyond the inventory
What the EU register of AI systems under Article 71 of the EU AI Act is, who is required, what information it needs and how to prepare it from governance.
-
EU AI Act compliance timeline: key dates after the Digital Omnibus
Updated EU AI Act timeline after the Digital Omnibus: which obligations apply and when, what is postponed to 2027-2028, and a downloadable PDF.
-
Emotionally intelligent AI in the enterprise: uses, risks and governance for Project Managers (2026)
Does AI feel? How emotionally intelligent AI works in the enterprise, its real uses, its risks and its governance under the EU AI Act, explained by an AI Project Manager.
-
Enterprise AI model evaluation: benchmarks, red teaming and metrics 2026
Enterprise AI model evaluation: six critical dimensions, public benchmarks vs your own golden set, red teaming and the metrics you should report to the committee.
-
Secure enterprise RAG: architecture, data leaks and the EU AI Act
Secure enterprise RAG: reference architecture, data-leak risks and EU AI Act compliance. A practical guide to taking Retrieval Augmented Generation to production without surprises.
-
EU AI Act August 2026: a compliance checklist for companies
EU AI Act August 2026 checklist: the 10 critical actions your company must complete before the deadline to avoid fines of up to €35M.
-
Can AI fill the cyber-expert deficit in banking?
AI cybersecurity banking DORA: why AI does not fill the cyber-expert deficit, it shifts it. An analysis of the real limit under DORA.
