Methods and practical sessions
Online lectures on machine learning, text and image data, econometrics, LLMs, agents, and AI-assisted software development.
ETH Zurich · Summer 2026
AI & Applied Economics
A two-week summer school for PhD researchers. We move from machine learning and multimodal data to large language models, agentic workflows, and applied research design.
Course overview
From research question to working application.
The course covers the full empirical pipeline: framing economic questions, constructing datasets with AI methods, building reliable software, and reasoning about identification and inference.
Research and project studio
In-person faculty sessions at ETH Zurich, with structured time for project development, feedback, and advisor meetings.
Week 1 · Online
Lecture schedule
31 August–4 September. All times are Central European Time. The Zoom link is available in the participant syllabus and course chat.
Monday 31 August
Course introductionElliott Ash
ML basics, embeddings & clusteringAndrea Ciccarone
Text & image dataAndrea Ciccarone
Tuesday 1 September
Machine learning & econometricsGermain Gauthier
Large language models & alignmentElliott Ash
Wednesday 2 September
Working with AI agentsFabian Roeben
Software development with AIFabian Roeben
Thursday 3 September
Training large language modelsFelix Drinkall
Open sessionTo be confirmed
Friday 4 September
Recap, conclusion & presentationsElliott Ash
Course library
Materials
Lecture slides will be posted here after instructors approve the final versions.
Course syllabus
Programme, faculty, preparation, assignments, and suggested readings.
Course introduction
Elliott Ash · Monday 13:00
ML basics, embeddings, and clustering
Andrea Ciccarone · Monday 14:30
Text and image data
Andrea Ciccarone · Monday 16:30
ML and econometrics
Germain Gauthier · Tuesday 14:00
LLMs and alignment
Elliott Ash · Tuesday 16:45
Working with AI agents
Fabian Roeben · Wednesday 14:00
Software development with AI
Fabian Roeben · Wednesday 16:15
Running LLMs for economic experiments
Felix Drinkall · Thursday 14:00
Week 1 assignment. Download the text-and-image problem set and the Fox/MSNBC YouTube dataset, or open the instructions and rebuild script on GitHub.
Problem set PDF → Dataset → Build script → Instructions → Open notebooks →Week 2 · ETH Zurich
Research sessions
7–10 September, RZ Building, Room F21. Mornings focus on faculty research. Afternoons combine sessions, project work, and advisor meetings.
| Monday · 7 Sep | 09:00–12:30 14:00–15:45 16:00–17:30 | Sergio Galletta Sergio Galletta Project studio |
|---|---|---|
| Tuesday · 8 Sep | 09:00–12:30 14:00–15:45 16:00–17:30 | Elliott Ash Philine Widmer Project studio |
| Wednesday · 9 Sep | 09:00–12:30 14:00–17:30 | Joachim Voth Project studio and advisor meetings |
| Thursday · 10 Sep | 09:00–12:30 14:00–17:30 | David Yanagizawa-Drott Project studio, consultations, and next steps |
People
Faculty
The summer school brings together economists and computational social scientists working across AI methods, policy, and institutions.
Before Week 1
Prepare your laptop
Complete these steps before the first online session so practical work can begin immediately.
Bring a laptop
Use a machine suitable for coding and browser-based notebooks.
Arrange AI access
Use ChatGPT Plus or Claude Pro, or contact the organizers about course-provided access.
Install an agent
Install Codex and/or Claude Code.
Install the basics
Set up Git and a working Python/Jupyter environment.