Diploma in AI Security at the Edge and Lightweight MLOps
Online postgraduate diploma of 800 hours in Edge AI Security and Lightweight MLOps: 8 modules covering threats to edge AI, model robustness and defence, on-device model protection, privacy and federated learning, lightweight MLOps, field monitoring and maintenance, and AI governance and compliance, with an assigned technical tutor.
Diploma in AI Security at the Edge and Lightweight MLOps — studying from Singapore
Singapore is one of Asia's leading electronics manufacturing centres, with semiconductor fabrication, back-end assembly and test, and precision electronics concentrated in areas such as Tampines and Jurong. The sector also includes embedded systems for industrial and automotive use and a strong telecommunications base led by the Infocomm Media Development Authority. Major employers include global semiconductor firms and electronics manufacturers. Engineers work on hardware design, firmware, communications systems and test engineering.
Fees for Singapore: S$1,560 (Singapore dollar), with instalments and employer sponsorship available.
Relevant bodies in Singapore
Typical roles
Questions from Singapore
How are electronics and telecom engineers recognised in Singapore, and does SEIUM count?
Telecom and electronics engineering is regulated by the Infocomm Media Development Authority (IMDA), and engineers may register with the Professional Engineers Board (PEB). SEIUM's qualifications are private, not state degrees, and not accredited by any national body. Whether they are recognised is decided by employers or the relevant body, so confirm this before enrolling.
Which Singapore employers hire embedded systems and telecom engineers?
Singapore has demand from telecom operators, semiconductor and electronics manufacturers, defence and aerospace firms, and industrial automation companies. Employers generally assess hands-on skills in firmware, hardware design and testing. Your final project can evidence embedded or telecom competence, but its value is for each employer to judge.
Can I follow this from Singapore while working, and does it need a visa?
Yes, the programmes are 100% online and taught in English, so no student visa is needed to study from home in Singapore. Expect a heavy workload: 1,900 hours for a master's, 800 for a postgraduate diploma and 300 for a specialist course. Fees can be paid in instalments, and employer sponsorship is possible.
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 specific threats to AI models deployed at the edge.
Apply defences against adversarial and poisoning attacks.
Protect on-device models through encryption and trusted environments.
Preserve privacy with federated learning and differential privacy.
Implement lightweight MLOps workflows for resource-constrained devices.
Monitor drift and maintain models in the field.
Document compliance with the EU AI Act and related standards.
Who it's for
- AI security engineers protecting models against adversarial and privacy attacks.
- MLOps engineers adapting pipelines to resource-constrained edge devices.
- Embedded AI engineers responsible for models running on cameras, gateways and sensors.
- AI risk analysts assessing threats and mitigations across the model lifecycle.
- AI compliance managers preparing documentation for the EU AI Act and ISO/IEC 42001.
- IoT platform engineers managing fleets of AI-enabled devices.
- Engineering or science graduates looking to enter the industrial IoT sector.
What you'll work with
- Regulation (EU) 2024/1689 – AI Act (harmonised rules on artificial intelligence)
- ISO/IEC 42001:2023 (AI management systems)
- NIST AI 100-1 (AI Risk Management Framework)
- ISO/IEC 23894:2023 (AI risk management guidance)
- NIST AI 100-2 E2023 (adversarial machine learning taxonomy)
- Regulation (EU) 2024/2847 (Cyber Resilience Act)
- IEC 62443-4-2:2019 (technical security requirements for industrial automation components)
- MLflow
- Edge Impulse
- Adversarial Robustness Toolbox (ART)
- AWS IoT Greengrass
- DVC
8 modules
1.1Attack surface of a deployed model1.2NIST taxonomy of adversarial attacks1.3Evasion with adversarial examples1.4Data poisoning and backdoors1.5Model extraction and theft1.6Model inversion and membership inference1.7Physical and side-channel attacks1.8Model supply chain risks1.9OWASP Machine Learning Top 10 list1.10Case study: threats to a smart camera
2.1Generating FGSM and PGD attacks2.2Adversarial training and its variants2.3Detection of anomalous inputs2.4Robustness certification2.5Defences against physical adversarial patches2.6Sanitisation of training data2.7Backdoor detection in networks2.8Evaluation with the Adversarial Robustness Toolbox2.9Trade-off between robustness and accuracy2.10Lab: attack and defence on a classifier
3.1Encryption of models at rest3.2Execution in trusted execution environments (TEE)3.3Model obfuscation and watermarking3.4Secure boot and remote attestation3.5Key protection in secure elements3.6Access control to inference3.7Side-channel mitigation3.8Model licensing and intellectual property3.9Integrity of the inference pipeline3.10Practical: encrypted model decrypted inside a TEE
4.1Differential privacy in training4.2Federated learning: FedAvg and variants4.3Secure aggregation of updates4.4Attacks on federated learning4.5Data minimisation and anonymisation4.6GDPR applied to sensor data4.7Assessment of privacy leakage4.8Flower and TensorFlow Federated frameworks4.9Communication and energy cost4.10Workshop: federated training with Flower
5.1Model lifecycle at the edge5.2Data and model versioning with DVC5.3Experiment tracking with MLflow5.4Model packaging and signed artefacts5.5OTA deployment of models5.6Edge Impulse and AWS IoT Greengrass platforms5.7CI/CD pipelines for embedded models5.8Automated model testing5.9Rollback and canary deployment5.10Lab: MLOps pipeline with MLflow and DVC
6.1Lightweight telemetry and inference metrics6.2Data and concept drift detection6.3Monitoring with limited resources6.4Selective data collection for retraining6.5Alerts and quality thresholds6.6Continuous retraining and validation6.7Management of heterogeneous fleets6.8AI incident response6.9Auditing and traceability of decisions6.10Practical: drift detector on an IoT device
7.1Risk classification under the AI Act7.2ISO/IEC 42001 management system7.3Risk management according to ISO/IEC 238947.4NIST AI RMF framework7.5Cyber Resilience Act requirements for AI7.6Technical documentation and model cards7.7Human oversight and transparency7.8Conformity assessment7.9SBOM and AIBOM for models7.10Case study: compliance file for a vision system
8.1Definition of the use case and device8.2Threat and risk model8.3Model hardening8.4On-device protection8.5Lightweight MLOps pipeline8.6Drift monitoring8.7Compliance documentation8.8System security testing8.9Report and presentation8.10Final project: final review with your tutor
How you'll learn
The programme is delivered 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 you can combine it with your job.
Each module combines theory, guided reading of the original standard and applied exercises. The working software is MLflow, Edge Impulse, Adversarial Robustness Toolbox (ART), AWS IoT Greengrass and DVC, in free, academic or evaluation versions. Testing is studied through real procedures: adversarial attacks on embedded classifiers, OTA deployment of signed models and drift monitoring in IoT fleets. Assessment is continuous, module by module, plus the final project.
Standards covered: Regulation (EU) 2024/1689 (AI Act), ISO/IEC 42001:2023, NIST AI 100-1 (AI Risk Management Framework), ISO/IEC 23894:2023, NIST AI 100-2 E2023 (Adversarial Machine Learning), Regulation (EU) 2024/2847 (Cyber Resilience Act) and IEC 62443-4-2:2019.
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 MLflow, Edge Impulse, Adversarial Robustness Toolbox (ART), AWS IoT Greengrass and DVC, 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 required), a computer able to run the programme software and English sufficient to read technical standards.
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
Postgraduate Diploma in IEC 60601/62304 and Clinical Risk
Online 800-hour Postgraduate Diploma in IEC 60601/62304 and Clinical Risk. 8 modules covering IEC 60601-1, IEC 62304 and Jama Connect. Assigned tutor and open enrolment.
Diploma in Deterministic Design and Ultra-Low Latency
Online postgraduate diploma of 800 hours in Deterministic Design and Ultra-Low Latency: 8 modules covering sources of latency and non-determinism, deterministic hardware architecture, low-latency operating systems…
Diploma in Hypervisors on ARM/x86 and I/O Virtualization
Online postgraduate diploma of 800 hours in hypervisors on ARM/x86 and I/O virtualisation. 8 modules covering virtualisation principles, ARM and x86 hardware virtualisation, I/O virtualisation, embedded…
More in Electronics, Embedded Systems & Telecom
All Electronics, Embedded Systems & Telecom programmes → · How to apply · Fees & financing · Scholarships