# Andreas Nissen

> Enterprise AI architecture, agent reliability, evaluation, and governed execution.

Senior Solutions Architect (AWS, Munich) working on agent reliability, evaluation, identity, policy controls, observability, and human approval for enterprise AI. This site is the canonical source for his writing and project evidence; every article and project below also has a clean Markdown version at the same URL plus `index.md`, and a machine-readable index of every page is published at https://andreasnissen.dev/index.json.

## About

- [About](https://andreasnissen.dev/about/): Background, focus areas, and contact links.

## Writing

- [Context and Control Are Different Layers](https://andreasnissen.dev/writing/context-and-control/): An agent needs enough context to reason and enough control to act safely. Treating them as one system hides different failure modes.
- [From Screenshots to Governed Tools](https://andreasnissen.dev/writing/from-screenshots-to-governed-tools/): Agent-friendly interfaces should expose structured operations without taking the human operator out of the loop.
- [An Ontology Is More Than Context](https://andreasnissen.dev/writing/ontology-is-more-than-context/): Context is the evidence available now. An ontology defines what the entities, relationships, and rules in that evidence mean.
- [Free AWS Builder Center Sandboxes Make Hands-On Learning Easier](https://andreasnissen.dev/writing/free-aws-builder-center-sandboxes/): AWS Builder Center removes the account and credit-card setup that often stops a workshop before the building begins.
- [The Hard Part of Agentic AI Starts After the Demo](https://andreasnissen.dev/writing/agentic-ai-after-the-demo/): A working prototype proves possibility. Production requires a separate discipline for autonomy, reliability, security, cost, and operations.

## Projects

- [Runbook Relay](https://andreasnissen.dev/projects/runbook-relay/): A deterministic incident-response control room where a human and an AI agent share evidence, approval state, and an audit log.
- [Mistral Reliability Lab](https://andreasnissen.dev/projects/mistral-playground/): An inspectable Python reference for bounded tool use, provider switching, resilience, observability, and deterministic evaluation.

## Work I Love

- [AI Agents: The Definitive Guide](https://andreasnissen.dev/work-i-love/ai-agents-definitive-guide/): A practical companion repository spanning agent architecture, production, evaluation, memory, cost, and security.

## Optional

- [RSS feed](https://andreasnissen.dev/index.xml): Subscribe to new writing.