Diploma in Drift Monitoring and Model Health
Online postgraduate diploma of 800 hours in drift monitoring and model health. 8 modules covering model degradation, statistical and sequential drift detection, label-free performance estimation, drift response and monitoring governance, with an assigned technical tutor.
Diploma in Drift Monitoring and Model Health — studying from the United States
Data science, artificial intelligence and simulation are central to American industry, with major concentrations in the San Francisco Bay Area, Seattle, New York and Boston. Engineering organizations increasingly rely on machine learning, digital twins and high-performance computing to accelerate design, testing and operations. Employers range from large technology companies to aerospace, automotive and energy firms that apply analytics to complex systems. Demand is strong for engineers who combine domain knowledge with modeling, statistics and software skills.
Fees for the United States: US$1,220 (US Dollars), with instalments and employer sponsorship available.
Relevant bodies in the United States
Typical roles
Questions from the United States
Which US bodies relate to data science, AI and simulation engineering?
There is no single licensing body for data science or AI in the US. Relevant professional organizations include the IEEE Computer Society, the Association for Computing Machinery (ACM), and the Society for Industrial and Applied Mathematics (SIAM). Federal guidance such as the NIST AI Risk Management Framework shapes practice. SEIUM's qualifications are private and not accredited by a national body, so recognition is decided by employers.
How do US employers in data and AI regard private online postgraduate qualifications?
US tech firms, financial institutions and research labs generally hire on the basis of portfolios, technical interviews, relevant experience and sometimes advanced degrees. A private online qualification can show structured study, but it is not a standard hiring requirement and each employer weighs it differently. Public projects, code repositories and measurable results tend to matter most.
Is it realistic to study data and AI from the US around a full-time job?
Yes, the programmes are fully online and taught in English, so you can study from home without a student visa. The workload is large: about 1,900 hours for a master's, 800 hours for a postgraduate diploma, and 300 hours for a specialist course. An assigned technical tutor supports you. Fees may be paid in instalments, and employer sponsorship is sometimes available, which you should confirm with your employer.
Take the controls.
Run a real engineering trade-off from this field. Move the sliders, watch the model react in real time and beat the challenge.
What you will be able to do
Identify the types and causes of model degradation.
Apply univariate and multivariate statistical drift tests.
Implement sequential drift detectors on data streams.
Estimate model performance without access to labels.
Build comprehensive model health indicators.
Respond in a structured way to drift incidents.
Govern monitoring according to the risk of each model.
Who it's for
- Machine learning engineers with models already running in production.
- MLOps engineers building monitoring and retraining pipelines.
- Data scientists who need to know when their models stop being reliable.
- Model quality managers setting health indicators and alert thresholds.
- AI risk analysts working on AI Act and ISO/IEC 42001 compliance.
- Model reliability engineers handling degradation incidents.
- Engineering or science graduates looking to move into manufacturing.
What you'll work with
- Regulation (EU) 2024/1689 (Artificial Intelligence Act)
- ISO/IEC 42001:2023 (AI management systems)
- ISO/IEC 23894:2023 (AI risk management)
- ISO/IEC TS 4213:2022 (assessment of classification performance)
- ISO/IEC 25059:2023 (quality model for AI systems)
- NIST AI RMF 1.0 (AI risk management framework)
- Evidently AI
- NannyML
- River
- Alibi Detect
- Grafana
8 modules
1.1Why models degrade1.2Covariate, label and concept drift1.3Gradual, sudden and recurring drift1.4Changes in sensors, recipes and raw materials1.5Feedback loops1.6Delayed and unavailable labels1.7Technical health versus statistical health1.8Taxonomy of silent failures1.9Real cases of industrial degradation1.10Workshop: drift risk catalogue for a model
2.1Kolmogorov-Smirnov test2.2Population stability index (PSI)2.3Kullback-Leibler and Jensen-Shannon divergences2.4Wasserstein distance2.5Chi-squared test for categorical variables2.6Maximum Mean Discrepancy2.7Correction for multiple comparisons2.8Window size and statistical power2.9Multivariate drift with domain classifiers2.10Practical exercise: comparison of drift detectors in Python
3.1Page-Hinkley and CUSUM3.2ADWIN and adaptive windows3.3Error-based DDM and EDDM3.4Stream detection with River3.5Drift detection in embeddings3.6Drift in time series3.7False alarms and detection time3.8Sensitivity tuning3.9Validation with synthetic drift3.10Lab: ADWIN on a sensor stream
4.1Confidence-based performance estimation (CBPE)4.2Direct loss estimation for regression4.3Probability calibration4.4Model uncertainty and conformal prediction4.5Out-of-distribution data4.6Business proxies for performance4.7Selective labelling and active sampling4.8Validation against delayed labels4.9Limitations of estimation4.10Practical exercise: performance estimation with NannyML
5.1Latency, error and availability metrics5.2Input data quality5.3Missing and out-of-range values5.4Feature integrity5.5Fairness and bias in production5.6Drift in SHAP explanations5.7Composite health indicators5.8Segmentation by plant, line and product5.9Health dashboards5.10Lab: health dashboard with Evidently and Grafana
6.1Decision tree when an alert fires6.2Diagnosing the cause of drift6.3Full or incremental retraining6.4Weighting of recent samples6.5Domain adaptation6.6Fallback models and safety rules6.7Communication with business users6.8Documentation of model incidents6.9Prevention in model design6.10Case study: drift after a change of raw material supplier
7.1Batch and online monitoring architecture7.2Storage of predictions and references7.3Tools: Evidently, NannyML, WhyLabs and Arize7.4Monitoring policies by criticality7.5Post-market monitoring requirements of the AI Act7.6Roles and responsibilities7.7Costs of monitoring7.8Audit and evidence7.9Maturity of the practice7.10Workshop: monitoring policy by risk level
8.1Selection of the model and its context8.2Drift risk analysis8.3Detector design8.4Performance estimation without labels8.5Health dashboard8.6Response procedure8.7Validation with simulated drift8.8Governance plan8.9Report and presentation8.10Final project: final review with the tutor
How you'll learn
Training is 100% online on the SEIUM campus, with downloadable material, recorded classes and an assigned technical tutor. There are no fixed timetables: you progress at your own pace within the intake, so the programme fits around your job.
Each module combines theory, guided reading of the original standard and applied exercises. The working software is Evidently AI, NannyML, River, Alibi Detect and Grafana, in free, academic or evaluation versions. Exercises are based on real procedures: simulation of synthetic drift to validate detectors, performance estimation without labels and analysis of model degradation incidents. Assessment is continuous by module, plus the final capstone project.
Standards covered: Regulation (EU) 2024/1689, ISO/IEC 42001:2023, ISO/IEC 23894:2023, ISO/IEC TS 4213:2022, ISO/IEC 25059:2023, NIST AI RMF 1.0.
Build it end to end
Your final project follows the real engineering workflow, step by step, reviewed by your tutor.
Where this takes you
Requirements
- Prior technical training or professional experience in the sector. No university degree is required.
- An internet connection and a computer capable of running the programme software.
- English sufficient to read technical standards.
Ready to start?
Send your application today. Our admissions team replies within 72 hours with your study plan and funding options.
Frequently asked questions
The programme totals 800 hours over 6–8 months. You set your own pace within each intake, and your tutor helps you plan the workload around your job.
You work with Evidently AI, NannyML, River, Alibi Detect and Grafana, using free, academic or evaluation versions. Your tutor indicates the free alternative wherever one exists.
A SEIUM Institute diploma, awarded on passing the continuous assessment and the final project. It is SEIUM's own qualification, not an official state degree; its recognition depends on the programme completed and the regulations of each country.
Yes. Modules passed are recognised in a higher programme in the same area (postgraduate diploma or master's). Ask about the recognition policy when you enrol.
Prior technical training or professional experience in the sector; no university degree is required.
Start dates are set by SEIUM's faculty for each intake. Our admissions team will confirm the next available date when you apply.
Related programmes
Diploma in Validation with Real Data and Gap Bridging
Online postgraduate diploma of 800 hours in real-data validation and bridging the gap. 8 modules covering the simulation-reality gap, real-data campaigns, domain gap measurement, domain adaptation,…
Diploma in Vision-Assisted GD&T and Dimensional Control
Online postgraduate diploma of 800 hours in vision-assisted GD&T and dimensional inspection. 8 modules covering the GD&T language, machine vision for metrology, geometric tolerance evaluation, in-line…
Diploma in Evaluation, Security and Observability in NLP
Online postgraduate diploma of 800 hours in evaluation, security and observability in NLP, structured in 8 modules covering language model evaluation, quality and bias, NLP application…
More in Data, AI, Simulation & Analytics
All Data, AI, Simulation & Analytics programmes → · How to apply · Fees & financing · Scholarships