Agentic Engineering · Software Architecture · AI Governance
This GitHub profile contains my own engineering projects, reference implementations, standards and tools.
My focus is on building reliable, auditable systems. That includes controlled tool access, explicit permissions, reproducible verification, human approval steps and traceable architecture decisions.
The repositories below cover agentic engineering, AI governance, RAG approval workflows, local agent pipelines and architecture decision tools. They document independent technical work, not deliverables from former client engagements.
Verification over claims.
- agentic-engineering-standard — how an engineering organization becomes agent-ready: a vendor-neutral standard with normative text, JSON Schemas, Rego policies and a runnable reference repository. Overview
- agentic-ai-governance-toolkit — how agents become governable: lifecycle models, risk scoring, EU AI Act & DORA checklists, and a working evaluator
- rag-approval-blueprint — how AI systems get approved: approval file, control catalogue with test procedures, deletion evidence, and a worked approval run (DORA/MaRisk)
- platform-decision-kit — how platform decisions get made: a triage & discovery instrument for brownfield systems — claims tagged as hypotheses, deliberately no recommendation. Try it in the browser — runs fully local.
- awesome-ai-governance — a curated list of regulations, standards, risk frameworks, and tools for governing AI systems
- local-agent-pipeline — a minimal, fully local multi-agent pipeline where auditability and human oversight are first-class features
- kickoff-ai — a reference setup for agentic development environments: staged installation, 41 verification checks, tool registry with sunset window
- razbiram-nlp — Bulgarian NLP enrichment engine with golden-set evaluated LLM prompts, powering razbiram.com
🌐 leon-koellerwirth.com · 🧰 Tools & working artifacts · 💼 LinkedIn