Perito IT supports teams and organisations in demanding IT projects. With deep expertise in software architecture, requirements engineering and modern AI-assisted development processes.
Agentic Software Engineering
AI assistants like Claude Code, Cursor or GitHub Copilot are changing how software is built — whether teams plan for it or not. The difference lies in whether their use is structured or haphazard.
I help teams build a solid foundation for Agentic Software Engineering: clear specifications as steering instruments for AI agents, review processes that keep pace with AI speed, and guardrails that prevent generated code from becoming a black box.
What this means in practice:
- Onboarding support — tool selection, workflow design and initial pilot projects
- Spec-Driven Workflows — establishing specifications as the source of truth for AI agents
- Quality Gates — review and test processes for agentic codebases
- Team Enablement — empowering developers to direct AI agents rather than follow them
Architecture
Thoughtful architecture concepts for complex systems — scalable, robust and future-proof. The connection between technology, business and user needs is always central: from high-level design through technology selection to documentation.
In the context of Agentic Software Engineering, good architecture today also means that systems remain readable and controllable for AI assistants. Clear module boundaries, explicit interfaces and structured decision documentation are not just good practice — they are a prerequisite for AI agents to operate meaningfully within a codebase.
- System architecture — for hardware, communication and software components
- Architecture reviews — technology evaluations and feasibility analyses
- Documentation — decision foundations and ADRs
- AI-Readiness — codebase structuring for agentic workflows
Requirements Engineering & Business Analysis
Clear requirements are the bottleneck in most projects — and at the same time the biggest lever. I support teams in eliciting, structuring and prioritising requirements for software solutions, cloud architectures and complex systems.
Precise requirements are also the key to effective Agentic Software Engineering: an AI agent that accesses a well-thought-out specification delivers predictable results. One that works on vague inputs produces vague output.
- Requirements engineering — elicitation, structuring and management of requirements
- Business analysis — process modelling and stakeholder analysis
- Specifications — user stories, use cases and acceptance criteria
- AI steering — specs as steering instruments for AI agents
CTO Advisory
Strategic programme design for technical leaders — not hands-on engineering coaching or implementation. Two advisory tracks for CTOs, VP Engineering, and CISOs.
AI Strategy & Governance
For organisations introducing AI-assisted development — structured, accountable, and compliant.
AI-generated code raises new accountability questions: Who is liable for a security flaw introduced by an agent? How is AI use documented and governed within the team? What nDSG implications arise when agents have access to personal data?
- AI strategy — Pragmatic adoption roadmap for Agentic Software Engineering within your team
- Governance & compliance — Accountability for AI-generated code, nDSG implications
- Technology transition — Experience guiding transformations in high-security environments and large organisations
Mythos-Ready Security
AI models now discover vulnerabilities faster than humans can close them. The CSA/SANS/OWASP “AI Vulnerability Storm” brief (April 2026) describes this as present reality — not a future scenario.
I give technical leaders a clear bridge between engineering practice and security strategy: spec-driven development and reviewer agents as practical security controls — not abstract governance.
- Board briefing — Situational assessment: what Mythos-class threats mean for your organisation
- VulnOps capability design — Building an operational vulnerability management capability
- Agentic SOC strategy — How a modern Security Operations Centre keeps pace with AI-speed threats
- Automated patch management — Strategy for automated patching at AI-driven threat density
- Secure-by-Spec — Spec-driven development as a security control: reviewer agents automatically validate AI-generated code against the spec
Interested? Contact me for a free 30-minute initial consultation.