Training Workshops on AI-Augmented Research Workflows for Social Scientists
On Tuesday, 7 April and Friday, 10 April, we will organize two full-day training workshops on the use of AI tools in social science research.
The workshops will take place at the ESSCA campus in Paris, from 10:00 am to 5:00 pm, and will be taught by Rubén Fernández-Fuertes (PhD candidate in Finance, Bocconi University) and Maksim Zubok (PhD candidate in Politics, University of Oxford), two leading early-career researchers with strong expertise in applying LLMs to social science research.
These hands-on workshops will show social scientists how to integrate AI tools into their daily research practice. Instead of just chatting with a chatbot, participants will learn to use agentic AI assistants that can read documents, write and execute code, search the web, and work together to tackle real research tasks: processing and OCR'ing academic papers, reviewing literature, cleaning datasets, running analyses, and drafting manuscripts. The session covers practical workflows using tools like Claude Code, OpenAI Codex, and Gemini, and shows how to combine them with knowledge management systems like Obsidian to build a personal AI-augmented research pipeline. The workshop is designed for researchers at all technical levels and progresses from accessible foundations to advanced multi-agent techniques, with hands-on exercises throughout. No prior programming experience is required.
Instructors
Rubén Fernández-Fuertes is a PhD candidate in Finance at Bocconi University and a former mathematician. His research focuses on macro-finance, monetary policy, and the application of large language models to financial economics. His job market paper develops a multi-agent LLM framework for analyzing Federal Reserve communications.
Maksim Zubok is a doctoral candidate in Politics at Oxford University specializing in political methodology. In his research, Maksim explores various ways of harnessing LLMs for social science research. Those range from classic data labelling to using the models as condensed snapshots of the internet from which we can learn how people organise relationships between concepts and thus form beliefs about the world they live in.