Alex P. Günsberg · Finance PhD Candidate, graduating Spring 2027 · Hanken · UNC Kenan‑Flagler
Finance PhD.
Data Scientist.
Agentic Engineer.
I study how households borrow and how lenders decide who gets credit, using causal inference on register data. I build software entirely with AI agents, and measure where they break and need a human.
On the 2026–27 job market · Open now to academic and industry roles
Academic research
more defaults within 18 months when applicants at the margin get the loan. For them, a rejection is protection.
loan applications from Finland's largest online loan marketplace: how people shop for credit when searching is almost free.
Agentic engineering
tasks shipped or merged since August 2026 by AI coding agents, with one person directing.
of their 1,555 code changes had to be undone after merging. Every break becomes a new check.
01 · Research
Who gets credit, and what happens next.
Household and consumer behavior in credit markets, studied with causal inference on register data.
- Protective Rejections in Consumer CreditJob market paper · November 2026
- Learning How To Borrow in a Fintech WorldNBER Working Paper 35024 · with Camelia M. Kuhnen
- The Path, Not the PositionWorking paper · solo-authored
02 · Industry & AI
Software built entirely by AI agents.
A training app built by AI coding agents, and a live dashboard that shows where they break and what fixes it.
- An AI-driven training appCase study · iOS and web
- Running a fleet of coding agentsCase study · up to 10 parallel lanes