YourNext
Opportunity.org

(MSc/PhD) AI Research Intern - LLMs, Causal Inference & Decision Making

Category
Internships
Posted
September 15, 2026
Region
Europe
Country
Netherlands
Sector
Technology & AI
Work mode
Hybrid
Apply via
jobs.ashbyhq.com

Overview

Prosus, a global technology company behind lifestyle e-commerce businesses such as Just Eat Takeaway, OLX and PayU, is offering an AI Research internship in Amsterdam for MSc and PhD students, centred on large language models, causal inference and decision making. Prosus spans more than 100 companies and has over 60,000 active AI agents helping teams work faster. Interns work alongside AI engineers who have published at NeurIPS and built production systems at scale.

What you will do

The intern builds and experiments with LLMs, representation models and modern machine learning systems for causal and decision-making problems, works with large-scale experimental data to see how users respond to different interventions, and builds models for treatment effects, uplift and counterfactual prediction. Other tasks are designing experiments, analysing randomised controlled trials, developing constrained optimisation models for promotion and incentive allocation, and reading research papers, prototyping ideas and turning the successful ones into real systems.

Who can apply

The internship is for students currently enrolled at a university who either hold EU citizenship or are studying in the Netherlands, since visas for internships cannot be supported. Candidates must be able to commit to six to twelve months in Amsterdam with at least three office days a week, and the team works around the academic calendar.

Requirements

Applicants should be strong in mathematics, probability and statistics, with solid quantitative modelling from machine learning, economics, finance, operations research or a similar field. They should understand causal and experimental reasoning, including randomised trials, counterfactuals, treatment effects and confounding, and be interested in optimisation under constraints. Good Python skills, comfort with neural networks, transformers and LLMs, and ideally PyTorch or JAX experience are expected. Experience with causal inference, econometrics, uplift modelling or mathematical optimisation is a plus, as is experience with promotions, pricing, ads, recommendations or marketplaces.

What you get

Interns access H200 GPU clusters, interaction logs from more than 500 million users and all major LLM APIs. The package includes competitive junior engineer compensation, a MacBook Pro, premium coding assistants, and a modern Amsterdam South office with free coffee and a travel card for commuting.