/ Research & Ideas

VeriBench: An End-to-End Formal Verification Benchmark for AI Coding Agents in Lean 4
VeriBench evaluates whether AI coding agents can turn Python programs into Lean 4 code with meaningful, machine-checkable correctness guarantees.
2026
↗

Research featured in a World Bank report: Co-worker, Coach, or Competitor?
A World Bank report features insights from my research with Yegor Denisov-Blanch on AI and software engineering productivity.
2025
↗

Measuring Determinism in Large Language Models for Software Code Review
Measuring Determinism in Large Language Models for Software Code Review
28.02.2025
↗

Research featured in The Washington Post: Are “ghost engineers” real?
The Washington Post covers our research into measuring developer productivity and “ghost engineers.”
December 2024
↗

Predicting Expert Evaluations in Software Code Reviews
Predicting Expert Evaluations in Software Code Reviews
September 2024
↗

Which country's software engineering teams perform the best?
March 2024
↗

The Great Engineering Reshuffle — Nationalmuseum, Stockholm
I presented The Great Engineering Reshuffle at Proxify Talks at Nationalmuseum in Stockholm on October 19, 2026.
↗

Keynote at adidas Tech Summit 2026 — Zaragoza
I delivered one of two external keynotes—alongside Gene Kim—at the 10th adidas Tech Summit in Zaragoza on October 1, 2026.
↗

From Vibes to Metrics: Why Evals Are Hard — AI Native DevCon London 2026
At AI Native DevCon London 2026, Rob Willoughby and I explained why output-only evals miss how AI agents work—and how trajectory evals and instrumentation make performance measurable.
↗

AWS Meeting of the Minds, Singapore: What actually works when building sustainable AI companies?
With global AI investments projected to reach $1.8 trillion by 2030, the panel explored the fine line between innovation and short-term hype.
↗

The 20% Truth: AI’s Real Impact on Developer Productivity
In this episode of Fractional, I discuss what the data really says about AI’s impact on developer productivity.
↗

Ghost Engineers with Yegor Denisov-Blanch and Simon Obstbaum
The conversation explores how different developers contribute varying value, how life circumstances impact work motivation.
↗

Roundtable on Engineering Excellence
A roundtable on engineering excellence, generative AI, and data-backed approaches to measuring software engineering productivity.
↗

Stanford Research: Measuring Software Engineering Productivity
Effective decision-making in software organizations relies on good data: you can’t improve what you don’t measure.
↗

McKinsey Roundtable: Software Engineering Excellence
An invitation-only McKinsey roundtable on practical approaches to software engineering productivity.
