Ezequiel
Lopez-Lopez
Research scientist · Senior AI consultant.
Measuring AI-mediated societal drifts.
- TU Dresden · SynoSys
- MPI for Human Development
- Freelance consultant
Research
Three foci. I work across their intersections. Each dot is a paper.
Since LLMs, new avenues to scale qualitative inquiry — particularly citizen-perspective elicitation at scale, combining LLMs and network analysis to surface and structure open-ended responses.
Catching generative AI in the act of reshaping language, cognition, and group behaviour. Observational studies of large media corpora (YouTube, podcasts, social media) for societal drifts and misinformation events; controlled behavioural experiments; LLMs used as instruments to study LLM effects.
Translating findings into evidence the policymaking process can use — in time, not after. Reporting, expert deliberation (Rapid Think Tanks), citizen elicitation on policy, brainstorming at scale.
Trajectory
Ten years across industry and academia.
-
2026 · expected PhD Computer Science TU Berlin · CS
-
2025 — now Research Associate TU Dresden · SynoSys + ScaDS.AI
2025 — now Guest Researcher MPI for Human Development · Berlin
Now: Dresden, Berlin, the next wave.
Research Associate at TU Dresden (SynoSys + ScaDS.AI); Guest Researcher at the MPI for Human Development, Berlin.
Large ongoing projects keep accumulating evidence that LLMs reshape spoken language — and the question is now scaling out. Identifying AI-driven societal drifts (linguistic, behavioural, cognitive, cultural), spotting the quasi-isolated environments where they can be measured cleanly, and combining macroscopic observation with controlled experiments on participants.
In parallel, mapping research directions and policy avenues to react in time — drawing on lessons from how slowly society absorbed the impacts of social media.
-
2022 — 2025 Doctoral Researcher · POLTOOLS MPIB · Center for Adaptive Rationality
POLTOOLS at MPI's ARC — where the PhD took shape.
Doctoral researcher at the MPI for Human Development's Center for Adaptive Rationality. Lead developer on the work packages of the DFG-funded POLTOOLS project — “Assisting behavioural science and evidence-based policy making using online machine tools.”
The original brief: build LLM- and network-based tools for citizen-perspective elicitation, argument mining, and behavioural-science methodology — informed by what the pandemic exposed about misinformation waves, restricted research-data access, and a sluggish science–policy loop. Also co-organised SciBeh's workshop cycle (2023–2025) as the community-facing arm of that work.
The packages then absorbed the GenAI wave, becoming the platform for the empirical-drift papers (linguistic drift, confirmation bias, hypercustomization, metacognition) and the method papers (citizen perspectives, Rapid Think Tanks). Published in Nature Human Behaviour, Behavioral Science & Policy, and the Annals of the NYAS.
-
2020 — 2022 Research Affiliate · NLP & IR Group UNED Madrid (remote)
2020 — 2022 AI & NLP Engineer DB Systel · Deutsche Bahn · Berlin
Two tracks at once: industry by day, research after hours.
Senior AI/NLP engineer at DB Systel (Deutsche Bahn IT), Berlin. Designed and shipped ML/NLP services for the digitalisation and automation of text-based processes across DB business units — document classification, NER, image segmentation, OCR pipelines, and information-retrieval systems running on internal infrastructure. Headline work: argument classification on customer correspondence (extracting structured complaint reasons from open-text letters) and an OCR-based document-intelligence pipeline for downstream classification and routing.
In parallel, External Research Affiliate at UNED's NLP & IR Group (Madrid, remote): focus on argument mining in the health domain, extending the MSc thesis work into computational measurement of controversy in non-political domains (drug-review forums, healthcare communities), with annotated datasets of user-review argumentation. The methodological foundation that fed the TU Berlin PhD.
-
2020 MSc, Language Technologies UNED Madrid
MSc · Language Technologies.
MSc in Language Technologies at UNED Madrid (online), specialisation in Natural Language Processing & Information Retrieval. Taken alongside the IBM Germany / DB Systel roles.
Coursework: Web Mining · Web Search Engines · Data Mining · Textual Information Discovery · Graph-Based Techniques Applied to NLP.
-
2016 — 2020 AI & Data Consultant IBM · Madrid & Berlin 2018 — 2020 · IBM Germany (Berlin) — banking, retail, insurance, media, automotive 2016 — 2018 · IBM Spain (Madrid) — banking, insurance, retail, energy, infrastructure
IBM · five years architecting AI from R&D to production.
IBM Spain (Madrid, 2016–2018) — first AI role. Among the first wave of production conversational systems deployed in Spain, across banking, insurance, retail, energy, and infrastructure on IBM Watson. Pre-LLM era meant the long version of NLP delivery: scoping, interviewing domain experts, restructuring fragmentary client knowledge bases that often weren't fit for purpose, designing dialogue flows / intents / IR layer end-to-end, then integrating with backend stacks (SQL, NoSQL, MongoDB, Elastic Search). This is where I learned to architect AI systems, not just write components.
Project work consolidated into four threads: conversational assistants (banking, retail, customer service); cognitive search & information retrieval (Watson Explorer + custom indexing); document classification and OCR pipelines (insurance, banking); and PoC / R&D explorations — IoT route optimisation for fleet operators, drone-based visual recognition, predictive analytics for assurance.
IBM Germany (Berlin, 2018–2020) — AI & Data Consultant. More document-classification pipelines, knowledge-base construction, and chatbot development — extended into the automotive sector alongside the existing banking, retail, insurance, and media mix.
-
2011 — 2016 BSc, Physics University of Granada
2015 Engineering Trainee European Space Agency · Madrid · Cluster Mission
Beginnings · physics, with an ESA traineeship.
BSc in Physics at the University of Granada, with a computational and mathematical-physics focus: statistical mechanics, electromagnetism, complex systems, computational physics, numerical methods, simulation, and programming.
Concurrently, an engineering traineeship at the European Space Agency (Madrid, 2015) on the Cluster Mission — first hands-on experience combining real measurement data with code.
That stack — analysis, statistics, signal processing, simulation, programming — later transferred cleanly into NLP, machine learning, and computational social science.
Work
A short bibliography. Full list in CV →
View all publications
In the press
Where the work has been written about.
Funded projects
Awarded funding I am named on, with role.
Organised workshops
Convening discussions across behavioural science, AI, and policy.
Talks
Recent first.
View earlier talks
Organisations
Scientific societies, networks, and platforms I'm part of.
- Coordinator SciBeh — Behavioural-science network for crisis knowledge management. Co-organiser of multiple SciBeh workshops (2023–2025) and Organising Committee member of the 2025 Anti-Autocracy Summit. scibeh.org →
- Support SaveScience.EU — Community wiki collating resources for scholars facing threats to science (support for imperilled researchers, IT security, open-science weaponisation). Companion to The Anti-Autocracy Handbook. wiki.savescience.eu →
- Member ACM — Association for Computing Machinery. International scientific society for computing. acm.org →
- Member DPG — Deutsche Physikalische Gesellschaft (German Physical Society). Aligned with the BSc Physics background. dpg-physik.de →