About

I am Pedram Madani. I build and ship AI in places where being wrong has real cost — mental healthcare at Levana, social services at Foreveryone, legal compliance at Legalithm.

Most of what is written about AI is wrong about what actually works in production. The technical writers are too close to the code; the strategy writers are too far from it. I sit in the middle — operator, researcher, four industries — and I write about what holds up under real conditions.

What I cover

What actually ships vs what demos well

Why AI adoption fails inside companies — and it is almost never the technology

Honest operator reads on the current AI cycle — what is real, what is trendslop

Research-to-production translation from my work at OVGU Magdeburg (published in IEEE, with five more papers in the 2026-27 pipeline)

Who this is for

If you are a CEO trying to figure out what your AI strategy should actually look like, or an investor trying to separate signal from slop, my essays are for you.

Receipts

CTO and co-founder at Levana / Health First GmbH — building mental-health AI in Germany. Tech lead for digital transformation at Foreveryone gUG — applying AI in the social sector. Founder of Legalithm — AI compliance for EU regulated industries. Researcher in Software Engineering and AI at OVGU Magdeburg, with one IEEE paper published and five in the 2026-27 pipeline.