ITS Prodigi is opening five courses for the 2026–2028 cycle. I handled the technical coordination and drafted the programmes of four of them: the two in Empoli and the two in Pisa.

The four programmes start from a concrete question: what does a developer or a security technician actually do now that tools like GitHub Copilot, Cursor and Claude Code have entered the daily working cycle. The value of our work has shifted from writing code by hand towards architectural design, systems and API integration, and the critical validation of what the AI produces. The Empoli courses build this competence — AI as a working tool — while the Pisa courses move up a level: designing, orchestrating and governing systems made of autonomous agents.

The four courses at a glance

  • Coding & AI Specialist (DEV26) — Empoli · national profile 10.1.1 Higher Technician, Software Developer · AI track
  • Cloud & AI Security Engineer (SYS26) — Empoli · profile 10.2.1 Higher Technician, System Administrator · Cyber track
  • Agentic AI Full-Stack Developer (FULLSTACK26) — Pisa · profile 10.1.1 Higher Technician, Software Developer · AI track
  • Agentic AI System & Cyber Security Expert (CYBER26) — Pisa · profile 10.2.2 Higher Technician, System Cybersecurity · Cyber track

Each course runs for 1,800 hours: 1,000 of classroom and lab work, 800 of company placement. Technical English up to B2, workplace safety and Agile project management are common to all four.

Empoli — AI at work

The two Empoli profiles use AI as a tool inside an already-defined line of work: one on software development, one on systems administration and security.

Coding & AI Specialist (DEV26)

Trains full-stack web and mobile developers who integrate AI into the development cycle. The stack runs from Linux and Git all the way to cloud deployment, and covers Java and Spring Boot, Python and Django, Node.js, PHP and Laravel, React and React Native. The modules AI: Prompt Engineering, Copilot and development tools and Python for AI: LLM, NLP and agents bring into daily work the professional use of generative AI, of large language models (LLM), of natural language processing (NLP) and of first software agents — including so-called “vibe coding”, describing in natural language what you want to build and collaborating with the AI to produce it. Security is cross-cutting — OWASP (Open Web Application Security Project), secure coding, GDPR — with a dedicated module on cybersecurity in the age of AI and an introduction to digital forensics. Every technical module produces a project work that goes into the student’s GitHub portfolio.

Cloud & AI Security Engineer (SYS26)

Trains system administrators and engineers of secure cloud infrastructure, with AI as a tool for automating security operations. It covers the full cycle: prevention (hardening, firewalls, network segmentation, risk assessment against standards such as ISO 27001, GDPR and the EU NIS2 directive), detection through a SOC (Security Operations Center) — SIEM (Security Information and Event Management), log analysis, OSINT (Open Source Intelligence) — and incident response following the NIST (National Institute of Standards and Technology) framework. On the infrastructure side: cloud on AWS, Azure and GCP with certification-track preparation, Docker and Kubernetes, continuous integration and delivery (CI/CD) pipelines with a DevSecOps approach, and scripting in Python, Bash and PowerShell for automation. The AI part is operational: behavioural analytics (UEBA — User and Entity Behavior Analytics), threat intelligence augmented by language models, response automation (SOAR — Security Orchestration, Automation and Response). Advanced digital forensics on mobile and cloud and a module on industrial and IoT/OT (Internet of Things and operational technology) security round out the profile.

Pisa — agentic AI

The two Pisa courses share a common premise: an agentic system is one in which one or more AI agents interact autonomously with APIs, databases and external services to complete a task. Designing and governing them calls for skills different from using an assistant.

Agentic AI Full-Stack Developer (FULLSTACK26)

Starts from a modern full-stack base (Python and FastAPI, TypeScript, React and Next.js, Node.js) and specialises it on AI: machine learning and LLM fundamentals, RAG (Retrieval-Augmented Generation) architectures with vector databases, and a substantial module — 68 hours, the largest of the AI-specific ones — on designing and building autonomous agents, with multi-agent orchestration through frameworks like LangChain and CrewAI. The exit profile is described as a team leader of AI agents: someone who sets the goals, assigns the tasks, and coordinates and supervises the work of autonomous agents the way a tech lead coordinates a team of developers. Security is specific to these systems: prompt injection, data poisoning, guardrails and keeping human control over decisions.

Agentic AI System & Cyber Security Expert (CYBER26)

Brings the same agentic approach into security, on both sides. On the offensive side: penetration testing with the PTES (Penetration Testing Execution Standard) methodology and the MITRE ATT&CK matrix of adversary tactics and techniques, ethical hacking with Metasploit, Nmap and Burp Suite, web vulnerability analysis (OWASP Top 10), OSINT and social engineering. On the defensive side: SOC, SIEM, NIST incident response and threat hunting. AI enters on both fronts — offensive AI (phishing generated by language models, adversarial machine learning, evasion of machine-learning-based antivirus, deepfakes) and defensive AI (UEBA, automated malware analysis, SOAR). Students design and supervise security agents: NLP-based honeypots, malicious-traffic classifiers, incident-response automation, multi-agent systems for threat hunting — in the role of someone who directs a team of agents the way a SOC manager coordinates a team of human analysts.

Two axes: national profile and degree of AI autonomy

The four courses read as a matrix. On one axis, two national profiles: the Software Developer (10.1.1) and the systems-and-security technician (10.2.1 in Empoli, 10.2.2 in Pisa). On the other, two degrees of AI autonomy. In Empoli, AI is a tool inside the working cycle: you use Copilot to develop, you automate a SOC with SOAR and UEBA. In Pisa, the object of the work becomes the agent itself — to design, orchestrate and supervise. The same phrase — team leader of AI agents — recurs in the descriptions of both Pisa courses, one on the development side and one on the security side, and it is the most compact way to point to where our work is heading.

What they share

  • Cybersecurity woven in across the board, not confined to a single course: every programme has a Cybersecurity in the age of AI module and one on digital forensics.
  • An individual project work in every technical module, published on the student’s GitHub profile and brought to the professional interview.
  • Technical English up to B2, Agile project management, and an Entering the Workforce module focused on the move into employment.
  • 800 hours of company placement out of 1,800 total, in territories — Empoli and Pisa — with a substantial ICT business fabric.

The full course descriptions and applications for the 2026–2028 cycle are at itsprodigi.it/corsi.

Wordmark: ITS Academy Prodigi.