FredrikSkauen
AI Engineer
Chief AI Officer at Prosper AI, Oslo
I build AI products with language models and decision models, and write about the ideas behind them in a way you don’t need a PhD to follow.
This is my working notebook: a place to document what I build, what breaks, and what I learn along the way, written down so the lessons outlast the projects.
Operethe work so far
- Sep 2025 – nowChief AI OfficerProsper AI, Oslo
Led AI development for Prospekt AI, acquired in a transaction valued at up to NOK 191M.
- Nov 2024 – Sep 2025AI EngineerProsper AI, Oslo
- May 2024 – Apr 2026Co-founder, Head of AIMetacog, Oslo
Co-founded Metacog, later acquired by Black Capital Technology. Built:
- Multi-agent candidate screening that reduces human bias
- Norwegian call transcription with automated QA and coaching
- Prompt design and agent workflows for AI-native companies
- Order-confirmation extraction into storage and ERP
- Clothing recolors and pose variations that keep the model’s identity
- Automatic screening and tagging of raw video footage
Attentionwhere mine goes
Most of it goes to two questions. What is actually happening inside these models, and how do you put them to work in production so they create real value, not just impressive demos?
- InterpretabilityLooking inside the model to understand why it does what it does.
- ProductionEvaluation, reliability and cost, the unglamorous parts that decide whether it works.
- ValueChoosing problems where a model changes the outcome, not just the interface.
Notes from the scriptorium
Explainers, field notes and essays on building with AI, all in one place.
Hello
Let’s talk
Building with AI, or thinking about it? I’m happy to share what I’ve learned, both what worked and what didn’t.