I build AI systems that work in production. Then I train the people who will run them after I leave.
Training, architecture, governance, Copilot agents, RAG, and AI security. English and Spanish.
Same weight. Different job. Sometimes you need the workshop. Sometimes the architecture. Sometimes the uncomfortable question in the room.
MCT-grade facilitation for technical teams, sellers, architects, and partners. Practical sessions. Real labs. No sleepy theatre.
Azure AI applications, Copilot agent design, governance frameworks, RAG, and security for AI systems. The fun part is making it boring enough to run.
Keynotes, panels, conference sessions, and advisory work on AI governance, agent accountability, and responsible AI. Clear opinions. Useful ones.
I have worked inside Microsoft, with Swiss telco partners, with training companies, with startup constraints, and with beta users who do not care about your architecture diagram.
Not claimed as delivered. Not buried either.
AI is not short of demos. It is short of ownership, evaluation, guardrails, and people willing to ask who gets the call when the agent gets it wrong.
Guardrails first. Features second. Yes, even when the demo looks amazing.
#EternallyCuriousMCT Summit North America, Redmond. Practical governance across data, interaction controls, privacy risk, and audit-ready oversight.
Microsoft AI Tour for Partners, Paris. The less glamorous layer that decides whether agentic AI becomes useful or just expensive.
Design the agent. Design the permissions. Design the rollback. Design the explanation. Then maybe let it touch the process.
Clear language for risk, privacy, security, and business value. Because a policy nobody can execute is basically office wallpaper.
I am Alejandro Araujo Rajzner. Based in Zurich, Switzerland.
I started as a telecommunications engineer. Then Microsoft. Then product work at Sunrise, running Microsoft and AWS partner portfolio topics in a Swiss telco context.
I also built a Swiss healthtech AI startup from the messy beginning. WordPress and Typebot prototype. Then FastAPI, PostgreSQL with pgvector, React Native, Azure Functions, Azure AD B2C, RAG, Switzerland North data residency, and quality evaluation across seven semantic dimensions.
Clinical advisory team. Beta users. Privacy constraints. Founder panic. The full menu.
Send the context. I will tell you if I can help. I may also tell you what not to build. Annoying, but usually cheaper.
