At the intersection of econometrics and machine learning — reconciling the two statistical cultures identified by Leo Breiman in 2001.
The core question: how do we move beyond prediction to causal estimation in high-dimensional settings? Double Machine Learning, causal forests, and double-robust estimators provide the methodological toolkit. Applications span labour markets, finance, energy, and regulatory economics.
Core module of the MPhil in Economics and Data Science. Causal inference meets machine learning — theory and applications in R and Python.
Machine Learning for Causal Inference. Proposal submitted to Cambridge University Press, April 2026.
The shift from AI as conversational tool to AI as autonomous actor — planning, executing, and coordinating across complex analytical workflows.
Commissioned by Ofcom's Economics and Analytics Group, this applied research demonstrates what agentic AI can deliver in a regulatory economics context: end-to-end analysis of 143,726 EU Digital Services Act moderation records, executed through multi-model orchestration with full reproducibility.
Agentic AI for Regulatory Workflows. DSA Statement of Reasons analysis — Discord Netherlands B.V., September 2024 to December 2025.
Perplexity Computer, Claude Cowork, Microsoft Copilot Studio, and Python–Streamlit: a comparative assessment for regulatory analytics.
What happens to human knowledge, scarcity, and wages when machines can learn? The central project — spanning academic paper, documentary film, journalism, and public lecture.
The Three Tiers framework provides the analytical backbone: AI does not dissolve Adam Smith's expertise-scarcity-wage chain — it relocates the scarcity node, from human expertise toward platform and infrastructure access. Presented at the Cambridge Festival of Ideas, Judge Business School, March 2026.
Target: Oxford Review of Economic Policy / Journal of Economic Perspectives. Three-tier taxonomy mapped to labour market outcomes.
Proposal in active development. The human story behind the economics — what happens to expertise when machines can learn it?
Teaching at the intersection of economics, machine learning, and artificial intelligence — from Cambridge MPhil modules to executive programmes and international summer schools.
Courses span the full range from graduate-level econometrics to practitioner-facing applied AI. The common thread is rigour: every course connects theory to real decisions, and every delegate or student leaves with methods they can use.
Causal Inference and Machine Learning. Core module of the MPhil in Economics and Data Science. Lent Term 2026.
Econometrics Summer School at Churchill College, Cambridge. Jointly taught with Professor Jeffrey Wooldridge. 20–24 July 2026.
AI for Strategic Decision-Making. Three-day residential at Møller Centre, Churchill College. Target launch: Autumn 2026.
Machine Learning for Causal Inference. Proposal submitted to Cambridge University Press, April 2026.
A Cambridge executive programme for board members and senior leaders — organised around the decision, not the technology. Three-day residential at Møller Centre, Churchill College.
Every session asks how AI changes a particular kind of executive choice. Delegates leave with six concrete artefacts — including a Causal Decision Canvas and a working AI agent — each tied to a real decision in their own organisation and presentable to the level above on Monday morning.
Not "interesting but vague." Five deliverables, each auditable, each tied to a real decision in the buyer's firm. Usable on Monday morning.
No peer programme in the executive AI market explicitly teaches causal reasoning to senior decision-makers. Pearl's Ladder is the intellectual spine.