Data Analyst Data Scientist Cybersecurity Analyst

Professional Data Scientist — From business problem to production model

A complete professional pathway for running a Data Science process from the business problem to an evaluated, deployed and monitored model.

A Data Scientist turns a business problem into an evidence-based experimental process. The role involves building reliable working datasets, formulating hypotheses, establishing baselines, developing and comparing models, measuring uncertainty and bias, and helping integrate models into products and operational processes.

This pathway covers Python, SQL, statistics, experimentation, Machine Learning, feature engineering, time series, text and embeddings, causality, deep learning, reproducibility, explainability, fairness and MLOps. The objective is to choose an appropriate modeling strategy, demonstrate both predictive performance and operational value, and contribute to deployment, monitoring and governance.

metiers_prof_data_scientist_outcome

15 courses • 150 interactive lessons • 45 to 75 hours of conversational learning

5 levels: Level 1 — 24,90 € excl. VAT Level 2 — 29,90 € excl. VAT Level 3 — 34,90 € excl. VAT Level 4 — 39,90 € excl. VAT Level 5 — 45,00 € excl. VAT

Preview of a lesson inside the Learning Workspace

Best offer

Complete Pack — Data Scientist

🎓 Certificate : "Professional Data Scientist" 523,80 €

📚 15 courses 📖 150 lessons ⏱ 45 to 75 h ⚡ 3000 credits included

399,00 € excl. VAT

Savings : 124,80 €

Statistical & Experimental Foundations Pack 3 courses Level 1

🎓 Certificate : Data Scientist — Statistical & Experimental Foundations

📚 3 courses 📖 30 lessons ⏱ 9 to 15 h ⚡ 480 credits included

74,70 €

64,90 excl. VAT

Savings : 9,80 €

01 Level 1

Problem framing and working environment

Understand the Data Scientist role, turn a business issue into a modeling problem, and set up a reproducible working environment.

Understand the Data Scientist role and distinguish analysis, prediction, causality, and optimization Turn a business challenge into a target variable, unit of analysis, and success criterion Set up a reproducible Python environment with notebooks, environments, and dependencies Work with arrays and data structures using NumPy and pandas Explore a dataset through distributions, cross-tabs, charts, and descriptive statistics Use Git to version a Data Science project from the exploration stage Visualize data with Matplotlib or Seaborn to communicate an initial reading Formulate testable business hypotheses before any modeling Organize a project into reusable folders, scripts, and notebooks Present an initial reading of the data to a business stakeholder

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 160 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

No advanced prerequisite. Everyday computer skills and an interest in data analysis or problem solving are sufficient to start.

Learning objectives

Understand the Data Scientist role and distinguish analysis, prediction, causality, and optimization
Work with arrays and data structures using NumPy and pandas
Visualize data with Matplotlib or Seaborn to communicate an initial reading
Present an initial reading of the data to a business stakeholder

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: Present an initial reading of the data to a business stakeholder
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

24,90 € excl. VAT

02 Level 1

Data quality and statistical reasoning

Identify data-quality issues and use the statistical and probabilistic foundations required for a Data Science project.

Identify missing data, outliers, measurement errors, and collection bias Understand samples, populations, variance, uncertainty, and confidence intervals Revisit conditional probability, independence, and Bayes’ theorem through business cases Understand covariance, correlation, and the limits of interpretation Prepare numerical, categorical, date, and simple text variables Test a statistical hypothesis and interpret a p-value without overusing it Understand common distributions (normal, Poisson, binomial) and their practical use Detect multivariate outliers Encode categorical variables (one-hot, ordinal, target encoding) according to context Audit the quality of a real dataset and produce a data quality report

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 160 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

No advanced prerequisite. Everyday computer skills and an interest in data analysis or problem solving are sufficient to start.

Learning objectives

Identify missing data, outliers, measurement errors, and collection bias
Understand covariance, correlation, and the limits of interpretation
Understand common distributions (normal, Poisson, binomial) and their practical use
Audit the quality of a real dataset and produce a data quality report

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: Audit the quality of a real dataset and produce a data quality report
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

24,90 € excl. VAT

03 Level 1

Experimental protocol and baseline

Build a rigorous experimental protocol and establish a first reference baseline before moving to more complex models.

Correctly separate training, validation, and test data Understand overfitting, underfitting, and the bias–variance trade-off Build a baseline before training complex models Choose an evaluation metric aligned with the business cost of errors Build a synthetic dataset to test a protocol before using real data Document an experiment plan before coding the first model Compare several naive baselines: mean, business rule, and random model Automate metric calculation and tracking in an experiment dashboard Justify the choice of a metric to a non-technical decision-maker Professional workshop: conduct a complete exploratory study and propose a modeling strategy

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 160 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

No advanced prerequisite. Everyday computer skills and an interest in data analysis or problem solving are sufficient to start.

Learning objectives

Correctly separate training, validation, and test data
Choose an evaluation metric aligned with the business cost of errors
Compare several naive baselines: mean, business rule, and random model
Professional workshop: conduct a complete exploratory study and propose a modeling strategy

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: conduct a complete exploratory study and propose a modeling strategy
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

24,90 € excl. VAT

Supervised Modeling & Validation Pack 3 courses Level 2

🎓 Certificate : Data Scientist — Supervised Modeling & Validation

📚 3 courses 📖 30 lessons ⏱ 9 to 15 h ⚡ 540 credits included

89,70 €

76,90 excl. VAT

Savings : 12,80 €

04 Level 2

Supervised learning experimentation protocol

Build a reliable supervised-learning experiment and choose a model family that supports the business decision.

Build a reproducible experimental protocol: data, transformations, baseline, metrics, and experiment log Choose between regression, classification, ranking, or probabilistic estimation based on the decision to support Build an interpretable supervised baseline and turn a probability into a decision rule Compare interpretable and nonlinear model families without confusing complexity with useful performance Understand ensembles, bagging, and boosting to improve robustness and generalization Choose an algorithm based on variable characteristics: linear, tree-based, or distance-based Understand L1/L2 regularization for controlling model complexity Build a robust, reusable scikit-learn pipeline Document a model’s assumptions and limitations from the design stage Compare a simple model and a complex model on the same concrete business case

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 180 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

A working knowledge of the foundations introduced in the first courses, or equivalent practical experience, is recommended.

Learning objectives

Build a reproducible experimental protocol: data, transformations, baseline, metrics, and experiment log
Compare interpretable and nonlinear model families without confusing complexity with useful performance
Understand L1/L2 regularization for controlling model complexity
Compare a simple model and a complex model on the same concrete business case

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: Compare a simple model and a complex model on the same concrete business case
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

29,90 € excl. VAT

05 Level 2

Data access and class imbalance management

Select a modeling strategy, access source data with SQL, and manage imbalanced classes and asymmetric error costs.

Select a modeling strategy based on volume, nonlinearity, interpretability, latency, and maintenance cost Query databases and data warehouses with SQL to build a working dataset Handle rare events, imbalanced classes, and asymmetric costs without manufacturing performance Choose metrics, curves, and decision thresholds based on the cost of false positives, false negatives, and abstentions Evaluate a regression model and analyze errors by segment, value range, and business use Join and aggregate data from multiple heterogeneous sources Oversample and undersample without leakage using SMOTE and alternatives Weight classes in the loss function instead of rebalancing the data Build a working dataset from near-real-time data streams Document the provenance and freshness of the data used (data lineage)

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 180 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

A working knowledge of the foundations introduced in the first courses, or equivalent practical experience, is recommended.

Learning objectives

Select a modeling strategy based on volume, nonlinearity, interpretability, latency, and maintenance cost
Choose metrics, curves, and decision thresholds based on the cost of false positives, false negatives, and abstentions
Oversample and undersample without leakage using SMOTE and alternatives
Document the provenance and freshness of the data used (data lineage)

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: Document the provenance and freshness of the data used (data lineage)
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

29,90 € excl. VAT

06 Level 2

Model validation and selection

Build robust validation, organize hyperparameter search, and select a final model using a defensible common protocol.

Build an appropriate validation strategy: holdout, cross-validation, groups, time, and dependencies between observations Organize hyperparameter search with an experiment budget, stopping criteria, and traceability Detect target leakage, temporal leakage, selection bias, and contamination between datasets Compare candidates under an identical protocol and document why a simpler model may be preferable Use Bayesian hyperparameter optimization to reduce compute cost Build a model ensemble (stacking, voting) and evaluate its real benefit Measure model stability through resampling (bootstrap) Choose between optimizing average performance and worst-case performance Write a model-selection note for peer review Professional workshop: run a scoring experiment, compare several model families, and defend the final choice

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 180 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

A working knowledge of the foundations introduced in the first courses, or equivalent practical experience, is recommended.

Learning objectives

Build an appropriate validation strategy: holdout, cross-validation, groups, time, and dependencies between observations
Compare candidates under an identical protocol and document why a simpler model may be preferable
Measure model stability through resampling (bootstrap)
Professional workshop: run a scoring experiment, compare several model families, and defend the final choice

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: run a scoring experiment, compare several model families, and defend the final choice
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

29,90 € excl. VAT

Representations, Time & Causality Pack 3 courses Level 3

🎓 Certificate : Data Scientist — Representations, Time & Causality

📚 3 courses 📖 30 lessons ⏱ 9 to 15 h ⚡ 600 credits included

104,70 €

89,90 excl. VAT

Savings : 14,80 €

07 Level 3

Feature engineering and introduction to deep learning

Design useful data representations and understand when deep learning adds value compared with classical Machine Learning.

Engineer relevant features from business knowledge without introducing leakage Reduce dimensionality and diagnose redundancy between variables Use PCA to explore high-dimensional data structures Build a K-means segmentation and assess its stability Introduce neural networks: perceptron, layers, backpropagation, and criteria for preferring deep learning over classical ML Build nonlinear interactions and transformations between variables Select features using filter, wrapper, and embedded methods Compare PCA, t-SNE, and UMAP for visual exploration of complex data Choose an activation function and loss function suited to the problem Train a small neural network with a modern framework and diagnose convergence

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 200 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

A working knowledge of the foundations introduced in the first courses, or equivalent practical experience, is recommended.

Learning objectives

Engineer relevant features from business knowledge without introducing leakage
Build a K-means segmentation and assess its stability
Select features using filter, wrapper, and embedded methods
Train a small neural network with a modern framework and diagnose convergence

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: Train a small neural network with a modern framework and diagnose convergence
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

34,90 € excl. VAT

08 Level 3

Time series and text data

Work with temporal and textual data using methods and validation strategies that respect their specific structure.

Analyze time series: trend, seasonality, autocorrelation, and structural breaks Build temporal validation without mixing past and future Create time-series features: lags, windows, aggregates, and calendar features Establish a forecasting baseline and compare simple models before complex ones Process text data through cleaning, representation, and simple supervised classification Model a time series with classical methods: ARIMA and exponential smoothing Handle multiple time series simultaneously with multi-series or hierarchical forecasting Detect regime shifts and changes in seasonality Represent text with TF-IDF and then pretrained embeddings Build an end-to-end text-classification pipeline for a business case

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 200 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

You should already master the pathway foundations or have equivalent experience. Advanced courses build on methods, tools and controls introduced earlier.

Learning objectives

Analyze time series: trend, seasonality, autocorrelation, and structural breaks
Establish a forecasting baseline and compare simple models before complex ones
Handle multiple time series simultaneously with multi-series or hierarchical forecasting
Build an end-to-end text-classification pipeline for a business case

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: Build an end-to-end text-classification pipeline for a business case
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

34,90 € excl. VAT

09 Level 3

Semantic representations, anomalies and causality

Use embeddings and anomaly detection, and distinguish prediction from causality when choosing an intervention.

Understand embeddings and their value for similarity and semantic representation Detect anomalies using statistical and algorithmic approaches Distinguish prediction from causality: an introduction to causal inference and uplift modeling for choosing an intervention Build an error strategy: edge cases, false positives, false negatives, and abstention Compare embeddings using cosine similarity and assess their quality Build an embedding-based similarity search system using nearest neighbors Detect anomalies with a simple autoencoder Estimate the causal effect of an intervention using matching or propensity-score methods Design an A/B experiment to validate a causal hypothesis Professional workshop: combine segmentation, prediction, and temporal analysis on a customer case

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 200 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

You should already master the pathway foundations or have equivalent experience. Advanced courses build on methods, tools and controls introduced earlier.

Learning objectives

Understand embeddings and their value for similarity and semantic representation
Build an error strategy: edge cases, false positives, false negatives, and abstention
Detect anomalies with a simple autoencoder
Professional workshop: combine segmentation, prediction, and temporal analysis on a customer case

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: combine segmentation, prediction, and temporal analysis on a customer case
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

34,90 € excl. VAT

ML Engineering & Production Pack 3 courses Level 4

🎓 Certificate : Data Scientist — ML Engineering & Production

📚 3 courses 📖 30 lessons ⏱ 9 to 15 h ⚡ 660 credits included

119,70 €

101,90 excl. VAT

Savings : 17,80 €

10 Level 4

Software engineering and ML reproducibility

Structure a Machine Learning project as testable, versioned, traceable and reproducible software.

Structure ML project code into testable, reproducible modules Version code, reference data, parameters, and model artifacts Track experiments: dataset, features, hyperparameters, metrics, and produced model Build an automated, rerunnable training pipeline Create unit tests for transformations, features, and scoring functions Use an experiment-tracking tool to compare runs Set up continuous integration for a Machine Learning project Manage dependencies and environments with reproducible tooling Document the code and internal API of an ML project for a team Review a peer’s code and provide constructive feedback on an ML pipeline

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 220 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

You should already master the pathway foundations or have equivalent experience. Advanced courses build on methods, tools and controls introduced earlier.

Learning objectives

Structure ML project code into testable, reproducible modules
Build an automated, rerunnable training pipeline
Set up continuous integration for a Machine Learning project
Review a peer’s code and provide constructive feedback on an ML pipeline

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: Review a peer’s code and provide constructive feedback on an ML pipeline
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

39,90 € excl. VAT

11 Level 4

Data reliability, explainability and fairness

Prevent training-serving discrepancies, explain model behavior, and assess fairness across relevant populations.

Validate schemas, prevent train–serve skew, and understand the role of a feature store Explain a model globally and locally using feature importance and SHAP-like methods Identify the limits of model explanations and avoid implicit causal claims Evaluate bias and performance across relevant subpopulations Build a prediction API and define an input/output contract Build a feature store shared by training and serving Compare LIME- and SHAP-style local explanations on the same case and interpret divergences Define context-appropriate fairness criteria: demographic parity and equal opportunity Mitigate a detected bias without excessively degrading overall performance Design non-regression tests for model quality and fairness

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 220 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

You should already master the pathway foundations or have equivalent experience. Advanced courses build on methods, tools and controls introduced earlier.

Learning objectives

Validate schemas, prevent train–serve skew, and understand the role of a feature store
Evaluate bias and performance across relevant subpopulations
Compare LIME- and SHAP-style local explanations on the same case and interpret divergences
Design non-regression tests for model quality and fairness

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: Design non-regression tests for model quality and fairness
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

39,90 € excl. VAT

12 Level 4

Deployment and model lifecycle

Deploy a scoring service and manage monitoring, retraining, rollback and model lifecycle in production.

Containerize or package a reproducible scoring service Set up service metrics: latency, errors, availability, and volume Monitor data drift, concept drift, and performance degradation Define retraining, rollback, and validation rules before promotion Deploy a model in batch and real-time modes and choose between them Set up progressive deployment using canary or shadow deployment Build actionable alerts instead of a passive dashboard Automate periodic retraining with validation before promotion Manage multiple model versions in production at the same time Professional workshop: industrialize a model from experimentation through monitoring

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 220 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

You should already master the pathway foundations or have equivalent experience. Advanced courses build on methods, tools and controls introduced earlier.

Learning objectives

Containerize or package a reproducible scoring service
Define retraining, rollback, and validation rules before promotion
Build actionable alerts instead of a passive dashboard
Professional workshop: industrialize a model from experimentation through monitoring

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: industrialize a model from experimentation through monitoring
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

39,90 € excl. VAT

Strategy, Governance & Foundation Models Pack 3 courses Level 5

🎓 Certificate : Data Scientist — Strategy, Governance & Foundation Models

📚 3 courses 📖 30 lessons ⏱ 9 to 15 h ⚡ 720 credits included

135,00 €

114,90 excl. VAT

Savings : 20,10 €

13 Level 5

Strategic framing and business value

Decide whether a problem truly requires Machine Learning and build the business case for a Data Science product.

Determine whether a problem truly warrants Machine Learning Build a business case linking model performance, operational impact, and economic value Define a production success metric distinct from offline validation metrics Choose between full automation, assisted recommendation, and human-in-the-loop operation Define responsibilities across Data Science, Data Engineering, IT, business, and governance Prioritize a portfolio of Data Science initiatives by value and feasibility Estimate the total cost of ownership of a Machine Learning product Negotiate a realistic project scope with a business sponsor Build a progressive roadmap from MVP to higher reliability Anticipate ML project failure scenarios and their consequences

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 240 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

You should already master the pathway foundations or have equivalent experience. Advanced courses build on methods, tools and controls introduced earlier.

Learning objectives

Determine whether a problem truly warrants Machine Learning
Choose between full automation, assisted recommendation, and human-in-the-loop operation
Estimate the total cost of ownership of a Machine Learning product
Anticipate ML project failure scenarios and their consequences

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: Anticipate ML project failure scenarios and their consequences
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

45,00 € excl. VAT

14 Level 5

Governance, ethics and documentation

Govern data and models, manage ethical and compliance risks, and document limits, responsibilities and authorized uses.

Build a production experimentation protocol and compare it with the current situation Manage privacy, security, and data-minimization constraints Identify risks of discrimination, proxy variables, and unfair decisions Document a model: scope, data, performance, limitations, prohibited uses, and dependencies Organize business validation and production-readiness criteria Set up a review committee for high-impact models Anonymize or pseudonymize sensitive data without destroying predictive value Manage consent and the rights of people affected by a model Audit an existing model after deployment to verify compliance Draft a responsible AI usage charter for a Data Science team

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 240 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

You should already master the pathway foundations or have equivalent experience. Advanced courses build on methods, tools and controls introduced earlier.

Learning objectives

Build a production experimentation protocol and compare it with the current situation
Document a model: scope, data, performance, limitations, prohibited uses, and dependencies
Anonymize or pseudonymize sensitive data without destroying predictive value
Draft a responsible AI usage charter for a Data Science team

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: Draft a responsible AI usage charter for a Data Science team
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

45,00 € excl. VAT

15 Level 5

Continuity, generative AI and executive communication

Secure the continuity of an ML product, integrate foundation models appropriately, and present the work to technical and business stakeholders.

Design a continuity plan: model failure, fallback, shutdown, and recovery Integrate foundation models into a Data Science product: prompt engineering, lightweight fine-tuning, RAG, and evaluation Present results to decision-makers without selling false precision Build an ML project portfolio with evidence of method, evaluation, and impact Evaluate the cost and latency of a foundation model versus an in-house trained model Build a robust RAG system with source and citation management Secure foundation-model usage against prompt injection and data leakage Combine a foundation model and a classical ML model in the same product Prepare an oral Data Science project defense for a mixed technical and business panel Final project: design, deliver, and defend a complete ML product with a production-readiness dossier

📚 10 lessons ⏱ 3 to 5 h 🎯 Quiz included ⚡ 240 credits included

Course information

Target audience

Data professionals, analysts, developers and technical profiles who want to carry out Data Science projects from exploration through evaluation, deployment and model monitoring.

Prerequisites

You should already master the pathway foundations or have equivalent experience. Advanced courses build on methods, tools and controls introduced earlier.

Learning objectives

Design a continuity plan: model failure, fallback, shutdown, and recovery
Build an ML project portfolio with evidence of method, evaluation, and impact
Secure foundation-model usage against prompt injection and data leakage
Final project: design, deliver, and defend a complete ML product with a production-readiness dossier

Target skills

Apply the course methods and tools to a representative professional situation.
Produce a usable and verifiable outcome by combining the concepts covered: design, deliver, and defend a complete ML product with a production-readiness dossier
Check the quality of data, results, evidence or decisions before communication or implementation.
Document the approach, explain limitations and defend the choices made to a professional stakeholder.

Estimated duration

3 to 5 h

Price

45,00 € excl. VAT