2026-09-10 –, Unconference
Over 10% of the global population cannot afford sufficient food, and over 40% cannot afford a nutritious diet. Enhance is a data platform that helps change that - by computing the cheapest, most nutritious, and most environmentally sustainable diets at national scale.
In this talk, we explore how the UN World Food Programme uses Enhance (a Python & AWS-based data platform), optimization and data science to support governments in making evidence-based food policy decisions. By modeling trade-offs between costs, nutrition (including micronutrient deficiencies such as iron and calcium), and environmental impact (CO₂ and water usage), Enhance enables policymakers to evaluate interventions such as subsidies, imports, crop replacements, and food fortification.
In this talk, we will showcase real-world applications of Enhance - already used in countries like Cambodia, Colombia and Chad - how Enhance-supports analyses contributes to agricultural policy decisions and helps improve access to more nutritious diets at scale for people across the globe.
Attendees will learn:
- How Enhance applies multi-objective optimization to real-world food systems
- How to model trade-offs between affordability, nutrition, and sustainability
- How data platforms can directly inform national policy
- Lessons from deploying data science solutions across multiple countries
An optional live demo will illustrate how optimization scenarios can be explored interactively.
This talk presents Enhance, a large-scale data platform developed by Capgemini, UN World Food Programme and Tilburg University's Zero Hunger Lab. Enhance combines operations research, data science, and policy modeling to improve global food systems.
Enhance solves a multi-objective optimization problem: identifying diets that minimize cost while meeting nutritional requirements and reducing environmental impact. The platform integrates diverse datasets—food composition, prices, environmental indicators, and demographic needs—and translates them into actionable policy insights.
The underlying models have been developed in collaboration with academic partners, including research groups from Tilburg University, with contributions from PhD researchers specializing in optimization and food systems.
We will cover:
Optimization approach: Linear and multi-objective programming to balance cost, nutrition, and sustainability
Data integration: Handling heterogeneous, country-specific datasets
Architecture: Python-based analytics stack deployed on cloud infrastructure
Policy simulation: Evaluating interventions such as subsidies, imports, and fortified foods
Real-world application: Case studies including Cambodia, where Enhance-informed insights supported agricultural policy decisions to improve dietary outcomes - like introducing new kinds of fortified rice as a result of analysis.
An optional live demo can showcase how optimization scenarios can be run interactively to explore policy trade-offs in real time.
Scope boundaries:
We focus on optimization modeling, system design, and applied impact. We do not cover deep learning architectures or low-level infrastructure tuning.
Target audience:
Data scientists, ML engineers, and researchers interested in applied optimization and real-world impact. Plus anyone who likes to use data to help feed the world better.
Required background knowledge:
-Basic Python
- Basic statistics
- Basic understanding of machine learning optimization concepts
Structure (30 minutes):
- Introduction: Global food challenge & role of Enhance — 5 min
- Modeling diets as multi-objective optimization problems — 10 min
- Platform architecture & data pipeline — 5 min
- Case study: Cambodia & policy impact — 5 min
- Optional live demo + Q&A — 5 min
As a data scientist, Marijn builds data-driven solutions to real-world problems at scale
As a social scientist, he understands how data translates into human behavior and policy impact
As an AI practitioner, Marijn develops and deploys systems that move beyond pilots into real-world use
As a speaker, he shares how data and AI can improve lives globally
Marijn combines technical expertise with a strong foundation in social science to design solutions that don’t just work in theory, but in practice. His work focuses on turning complex data and optimization models into actionable insights for decision-makers.
Over the past years, he has been closely involved in the development of large-scale data platforms such as Enhance, applying multi-objective optimization to help governments design affordable, nutritious, and sustainable food systems.
At Capgemini, Marijn has spent over 9 years leading AI and data science initiatives—from early-stage exploration to production-grade deployments with real-world impact.
He is a frequent speaker on AI, data science, and the gap between experimentation and meaningful adoption.
Most of all, Marijn is driven by one goal: using data to improve people’s lives at scale.