I help engineering teams ship AI past the demo, into systems customers depend on, budgets can sustain, and engineers can still understand six months later.

Most AI advice is written for greenfield builds and demo loops. The teams I work with have history: existing codebases, real users, finite budgets, and senior engineers who've seen this movie before. The hard work is everything after the demo: evals that aren't vibes, agent architectures that survive contact with production, FinOps discipline that keeps the bill survivable, and code review when 80% of code is AI-generated.

20 years building production systems at banks (UBS, Barclays), fintechs, and startups. Previously at Accenture, IBM, Sun Microsystems, the Financial Times, and BBC. The common thread: environments where breakage has consequences.

I write about this at The Principal Engineer, twice a week.

Languages: fluent English, Swedish, Finnish & professionally proficient German (C1).