Universities and schools
Master’s programs, engineering schools, university diplomas, doctoral training, and specialized courses.
Upcoming masterclass
Understanding and modeling patient trajectories
Training · Higher education · R&D teams
An advanced, modular training offer covering temporal medical data, clinical AI, explainability, and neuro-symbolic approaches.
Who is it for?
The content, mathematical depth, and practical cases are adapted to the audience, learning objectives, and available time.
Master’s programs, engineering schools, university diplomas, doctoral training, and specialized courses.
Data, AI, innovation, product, and research teams addressing the specific challenges of healthcare AI.
Innovation and clinical research teams that need to work effectively with AI system designers.
Guest teaching, module design, or integration of the curriculum into an existing learning offer.
Learning outcomes
The goal is not to accumulate tools. It is to build sound judgment for designing, evaluating, and discussing healthcare AI systems.
Define your learning objectivesIdentify the specific challenges of longitudinal and temporal medical data.
Select a model family suited to the clinical question and project maturity.
Understand statistical, neural, and hybrid models before using them.
Assess performance, uncertainty, bias, and the limitations of predictive systems.
Design an explainability approach aligned with clinical use.
Improve collaboration between technical teams, researchers, and healthcare professionals.
Available formats
Durations are indicative. Each proposal is built around your audience, constraints, and expected outcomes.
1.5 — 3 hours
Introduce a cohort, team, or leadership group to a focused healthcare AI topic.
1 — 2 days
Combine concepts, demonstrations, and hands-on work around a use case or business challenge.
8 — 16 hours
Explore temporal data, clinical AI, explainability, critical care, or neuro-symbolic methods in depth.
20 — 60 hours
Build a progression tailored to a degree program, R&D team, or professional upskilling initiative.
100 hours
Deliver the ten-module reference curriculum, from temporal modeling foundations to advanced clinical systems.
Share the audience, objectives, and expected duration, and I will propose a coherent format.
Reference curriculum
The full curriculum represents 100 hours. It acts as a learning library: modules can be selected, shortened, combined, or extended according to participant level.
This module introduces the mathematical foundations required for modern temporal modeling. Participants develop a rigorous understanding of stochastic processes, filtrations, temporal information flow, and dynamical systems. The objective is to build a strong mathematical intuition for advanced temporal AI systems used in healthcare settings.
Probability spaces • Random processes • Stochastic processes • Filtrations • Adapted processes • Information flow • Temporal observability • Martingales intuition • Dynamical systems basics • Temporal operators
This module covers the major statistical approaches used in time series analysis. Participants learn how to model temporal dependencies, trends, uncertainty, and latent temporal states. This module provides the statistical backbone required before moving toward deep learning approaches.
Stationarity • AR models • MA models • ARIMA / SARIMA • State-space models • Kalman filtering • Hidden Markov Models (HMM) • Gaussian Processes • Forecasting theory • Uncertainty estimation
Introduction to real-world healthcare temporal data and Intensive Care Unit (ICU) ecosystems. Participants discover the structure and complexity of hospital data pipelines, physiological signals, and clinical event streams. Special attention is given to irregularity, missingness, and clinical operational constraints.
ICU data ecosystems • Electronic Health Records (EHR) • Clinical event streams • Physiological waveforms • Irregular sampling • Missing values • Censoring • Temporal alignment • Clinical coding systems • Healthcare data pipelines
This module introduces signal processing methods applied to physiological and biomedical data such as ECG and EEG. Participants learn how to preprocess, analyze, and extract meaningful temporal information from noisy clinical signals.
Fourier analysis • Spectral representations • Filtering techniques • Signal denoising • Wavelets • Temporal feature extraction • ECG preprocessing • EEG preprocessing • Multiscale temporal analysis
Advanced introduction to deep learning architectures designed for temporal and sequential data. Participants learn how modern neural architectures capture complex patient trajectories and long-range temporal dynamics.
Recurrent Neural Networks (RNN) • LSTM • GRU • Seq2Seq architectures • Temporal CNNs • Attention mechanisms • Transformers • Time embeddings • Temporal representation learning
Application of temporal AI systems to real clinical prediction tasks including mortality prediction, deterioration detection, and patient trajectory forecasting. Participants explore realistic ICU prediction workflows.
Early warning systems • Mortality prediction • Sepsis prediction • Clinical deterioration detection • Patient trajectory forecasting • Risk scoring systems • Longitudinal patient modeling • ICU outcome prediction
This module focuses on interpretability, transparency, and trust in clinical AI systems. Participants discover methods used to explain temporal models and ensure responsible deployment in healthcare environments.
Explainable AI (XAI) • Temporal saliency • Attention interpretability • SHAP for time series • Counterfactual explanations • Clinical interpretability • Fairness and bias • Trustworthy AI systems
Introduction to causal inference and temporal reasoning in clinical systems. The module explores the difference between correlation and causality in patient trajectories and medical interventions.
Temporal causality • Counterfactual reasoning • Treatment effect estimation • Causal graphs • Time-dependent confounding • Sequential decision modeling • Clinical interventions • Dynamic treatment analysis
Advanced module dedicated to temporal logic, symbolic reasoning, and neuro-symbolic AI systems. Participants learn how symbolic temporal constraints can be combined with neural learning systems for explainable clinical AI.
Temporal logic • Event calculus • Symbolic temporal reasoning • Neuro-symbolic AI • Hybrid clinical systems • Knowledge-guided learning • Rule-based temporal inference • Explainable symbolic constraints
Exploration of modern research directions in temporal AI for healthcare including foundation models, self-supervised learning, and multimodal clinical systems.
Foundation models for healthcare • Self-supervised learning • Contrastive temporal learning • Multimodal clinical AI • Large clinical models • Representation learning • Modern research directions • Open research problems
Specialized tracks
These tracks provide a starting point. They can be adapted to your level, sector, and organizational objectives.
Audience: Data Scientists, AI Engineers, Master Students
Audience: Machine Learning Engineers, AI Researchers, MedTech Professionals
Audience: ICU Researchers, Clinical AI Engineers, Healthcare Innovation Teams
Audience: AI Researchers, AI Governance Teams, Responsible AI Specialists
Audience: PhD Students, Advanced Researchers, Neuro-Symbolic AI Specialists
Teaching approach
Why this training?
Practical information
Let’s discuss your training project
Share the context, participant count, expected level, preferred format, and timeline. You do not need a complete specification to start the conversation.