Accessful GmbH
AI in production
Co-founder, Managing Director and Co-CTO
Platform in production use with clients from automotive, insurance, financial services and public administration.
Enterprise AI Architect
Dejan Guberinic
Agentic AI & Software Development Transformation
I help large organisations to deploy generative AI and AI agents safely and productively — particularly in software development and knowledge work. Focus: enterprise architecture, agentic SDLC, context engineering, governance, security, adoption and scaling AI across teams.
Focus enterprise agentic AI transformation, and within that agentic software development in particular. Not “expert on everything to do with AI”.
Three pieces of evidence, not three claims
Every statement in this portal has a source — a reference, a certificate, a repository or a passed check-in. What has no source is not in here.
SIGNAL IDUNA Group · self-declaration
Agentic SDLC inside a corporate
- Introduction and further development of agentic software development within the company.
- Contribution to a leading technical team responsible for planning, coordinating and delivering a proof of concept for agentic software development.
- Production of playbooks, ways of working and guidelines for AI-assisted and agentic development practices.
iSAQB
Architecture with proof
CPSA Foundation Level — Examination certificate (passed), Examination 26.04.2023, Hamburg.
T 20 · M 30 · C 40 = 90
The number nobody else shows
On 10.08.2026 the domain knowledge for the target role was measured across 20 knowledge questions and 5 architecture cases — no aids, assessed by an agent instructed not to flatter.
2.65 / 4 Weighted overall score
3 / 4 Target
3 Areas at level 2 — of 7 · 40 % of the weighting combined
Target not yet reached
The score is public because it will be measured again. Every two weeks the agent tests against demonstrated answers and records the result — upwards or downwards. A level nobody can evidence is not shown here.
This portal maintains itself
Content, CV and learning status live as Markdown in a Git repository. An AI agent updates them on a two-week cycle, every change is a commit, and the website is generated from it. The change history is public.
See the full picture
Skill matrix with evidence, career history and references, projects, learning path, CV and terms sit behind a personal access link — one per recipient, revocable, with an expiry date.