Luka Breitig

Luka Breitig is an AI Agent Trainer and Software Engineer: an engineer who studied people. He trains engineering and marketing teams to work with AI agents, so they get more done with the people they already have, and he builds the software when a team needs it built. He works in English and German, on site or remotely.

His background is unusual for an engineer. He holds a B.Sc. in Psychology and an M.A. in Management, and is a triple scholarship recipient, including the Stiftung der Deutschen Wirtschaft (SDW), placing in the top 2% nationally. That is why he treats AI adoption as a people problem as much as a software one: an agent only scales a team's work once people trust it and change how they work.

Before this, Luka spent six years as Co-Founder & CTO of the marketing agency The Happy Beavers, where he automated 80% of the business processes. Working next to marketers every day showed him where automation pays off and where human judgement still decides the result.

Today his work covers three equal areas:

• Training: engineering and marketing teams on AI agents, Claude Code, Cursor, Codex and MCP servers
• Implementation consulting: getting AI agents adopted in real workflows
• Development: agent systems and custom software in Elixir/Phoenix and Python

He delivers training commissioned by L&D providers including Cegos Integrata and NobleProg, has spoken at international conferences on AI integration, and designs custom programmes for teams bringing agents into their existing work.

His notable projects include:

• Tymeslot - self-hostable scheduling platform with real-time calendar sync, embeddable booking widgets and n8n integration
• Transvexis - multilingual translation platform serving clients in over 39 languages
• In A Nutshell - AI semantic search for video content with timestamp citations

Luka believes the best technical solution is the one a team actually uses. His writing covers practical patterns from training teams, building AI products and putting LLMs into real applications, all drawn from systems he has built and run himself.