Perito IT GmbH
Agentic Software Engineering:
Systematic. Secure. Controllable.
AI agents accelerate development — but they increase complexity and security risk. We bring the methodical rigour of critical infrastructure engineering to your AI-assisted software lifecycle.
Methodological foundation: Taproot — open-source framework for Spec-Driven Development
Why Perito?
Critical Infrastructure
Experience with high-availability systems in transport — where reliability is not a feature, it's a prerequisite.
Software Security
10 years of experience in software security and technical leadership — knowing attack vectors means building better architectures.
CTO Perspective
Strategic guidance through technology transitions — we speak the language of business and understand the liability risks of AI transformation.
Current courses
Hands-on training for engineering teams — on-site or remote, immediately applicable.
Spec Driven Design with AI
From specification to code — methodical and AI-assisted
- Use specs as source of truth for AI assistants
- Hands-on with OpenSpec, Claude Code and Cursor
- Integration into existing processes (Brownfield)
Secure Agentic Coding
Deploy AI agents safely in production environments
- Understand OWASP relevance for AI-generated code
- Secure agent outputs and build automated security gates
- Prompt engineering for security-critical architectures
Custom in-house training — get in touch
Consulting & project support
Perito IT supports teams and organisations in demanding IT projects — from the first requirement to go-live. With deep experience in software architecture, requirements engineering and project management.
- Architecture
Well-considered concepts for complex systems — scalable, robust, future-proof.
- Requirements Engineering
Clear, realisable requirements — technical understanding with a view of the whole.
- Project management
Structure, pace and clear communication — from kick-off to go-live.
For decision-makers
CTO Advisory: AI Strategy, Governance and nDSG Compliance
Technical leaders carry responsibility for AI-generated code — but few frameworks provide answers. We help you develop a pragmatic AI strategy, build governance structures and clarify the nDSG implications of agents — before your legal department asks.
Your contact
Stefan Aeschbacher
Engineer with over 25 years of IT experience, including 10 years in software security. Background in critical infrastructure (railway technology, PostAuto) and technical leadership. Today: Agentic Software Engineering and AI governance.