BiGHT - Course Overview

Building intelligent Global Health and Humanitarian Response Technologies (BiGHT)

Step 1 of 5

Summary

CS-467 Course Curriculum - Fall 2026 - EPFL IC

Summary

This course teaches the principles and practice of designing, building, and rigorously evaluating trustworthy, representative, and contextually appropriate AI systems for high-stakes decision making in global health and humanitarian response.

Lecturers: Prof. Mary-Anne Hartley, Prof. Kristina Keitel, Dr. Lars Klein, Dr. David Sasu, Tim Arni, Yusuf Kesmen, Fabrice Nemo

Coordinator: Fabrice Nemo

TAs: Yusuf Kesmen, Xavier Theimer-Lienhard, Jérémy Baffou

Course Overview

The course is organized into three thematic blocks. Each teaching week pairs one context lecture with one engineering lecture.

Block	Weeks	Goal
Volatile Contexts	1-4	Understand the environments in which AI must operate, and why technology often fails in humanitarian and clinical settings.
High-stakes Decisions	5-7	Understand how clinicians make decisions under uncertainty and how AI can safely support human expertise.
Trustworthy Evidence	9-13	Design, evaluate, and deploy AI systems that are safe, effective, and ready for real-world use.

Week	Theme	Context lecture	Engineering lecture
1	Inequitable Inaccuracy	Tutti Fratelli: The Principles That Created a Movement	Building for Broken Environments
2	When the Hospital Is the Target	When the Hospital Is the Target	Systems That Work When Nothing Else Does
3	Listening Before the Outbreak	The Map That Stopped an Epidemic	Listening to the World: Epidemic Intelligence at Scale
4	The Last Mile	Health for All, Care for Few	The Last Mile: Getting Intelligence to the Edge
5	Models Without Patients	The Machine That Was Right and Was Never Used	From Rules to Reasoning
6	Reasoning Against Ourselves	The Confident Fool: How Expertise Breeds Error	Debiasing the Machine
7	Advice at the Wrong Time	Why Doctors Ignore Good Advice	Integration Without Interruption
8	Project Studio and Midterm Review	Midterm presentations / project checkpoint	Technical design reviews
9	The Fragility of Accuracy	How We Learned to Count What We Were Losing	When 99% Accuracy Is Dangerous
10	Causal Evidence in Practice	The Man Who Asked Medicine to Prove Itself	Causal Inference: Beyond Correlation
11	Designing for Failure	The Holes in the Cheese	Uncertainty as Architecture
12	Responsibility in Clinical AI	First, Do No Harm - Then, Prove It	Governing What You Cannot Fully Understand
13	Systems That Reach People	Why Good Ideas Die at the Door	Shipping to Scale: The MLOps of Global Health AI
14	Final Presentations	Final Presentations	Final Presentations

Thematic Blocks

Block I: Volatile Contexts (Weeks 1-4)

Goal: Understand the environments in which AI must operate, and why technology often fails in humanitarian and clinical settings.

Context themes

Humanitarian systems
Community medicine
Global health
Epidemics and outbreaks

Engineering themes

Participatory design
ML infrastructure
Federated learning
Privacy
Reproducibility
Edge computing

Block II: High-stakes Decisions (Weeks 5-7)

Goal: Understand how clinicians make decisions under uncertainty and how AI can safely support human expertise.

Context themes

Clinical reasoning
Diagnosis and triage
Evidence-based medicine
Clinical workflows

Engineering themes

Data analysis
Bayesian reasoning
Causal inference
RAG
Clinical decision support

Block III: Trustworthy Evidence (Weeks 9-13)

Goal: Design, evaluate, and deploy AI systems that are safe, effective, and ready for real-world use.

Context themes

Clinical trials
Implementation science
Regulation and governance
Health systems

Engineering themes

Validation
Benchmarking
Safety
Deployment
Monitoring
Technical communication

Organization

Item	Description
Weeks	14
Lectures	24 paired lectures: 12 context lectures and 12 engineering lectures
Project studios	12 studio sessions with the team to help you through your projects
Weekly workload	2h lectures, 1h project studio, ~12h project development
Presentations	Midterm and final project presentations
Quizzes	6 in-class quiz sessions
Field trip	Optional fully funded field trip to the ICRC museum
Final output	A deployable AI system for Global Health / Humanitarian Response and technical documentation

Weekly format

Session	Purpose
Context lecture	What is the context of the problem we are solving?
Engineering lecture	Translate context concepts into engineering principles that address them
Studio	Apply the week's ideas to the semester-long project.

Assessment

Component	Weight
Milestone 1: Technical Design and Reproducible Repository	10%
Milestone 2: Proof Of Concept	20%
Midterm oral presentation	10%
Final presentation with a live demo	10 %
Project report and code	30%
In-class tests and quizzes	20%