Master’s in Edge AI & TinyML
Online master's programme of 1,900 hours in Edge AI and TinyML: quantisation and pruning, runtimes on MCUs and NPUs, and secure OTA updates. 12 modules covering hardware, on-device data, efficient models, compression, deployment frameworks, accelerators, fleet operation and security, with an assigned technical tutor.
Master’s in Edge AI & TinyML — studying from Australia
Data, artificial intelligence and simulation are central to Australian industries including mining, banking, health, agriculture and government. Organisations use machine learning, digital twins and advanced analytics to improve productivity, asset performance and decision making. National science agency CSIRO, along with universities and a growing start-up ecosystem, drives research and applied innovation. Australia has also developed national AI ethics principles, making responsible design an important professional consideration. Engineers in this area typically work for technology firms, consultancies, research organisations and large enterprises.
Fees for Australia: A$12,600 (AUD), with instalments and employer sponsorship available.
Relevant bodies in Australia
Typical roles
Questions from Australia
Which Australian bodies recognise data and AI engineering skills?
Engineers Australia and the Australian Computer Society (ACS) are the main professional bodies, and the ACS runs a skills assessment for migration purposes. SEIUM's private qualifications are not accredited by either. Whether a private qualification is valued is decided by the employer or the relevant body.
What data and AI careers exist in Australia for this training?
Australia has demand for data scientists, AI engineers and simulation analysts across finance, health, mining, government and technology firms. Employers mainly assess your portfolio, technical skills and results. A SEIUM qualification may support your profile, but each employer decides its relevance.
Can I study this online from Australia while working?
Yes, the programmes are fully online and taught in English, and no student visa is needed to study from home. Specialist courses are 300 hours, postgraduate diplomas 800 and master's 1,900, with an assigned technical tutor and a final project. Fees can be paid in instalments, and employer sponsorship may be 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
Select suitable hardware for inference at the edge.
Capture and preprocess sensor data on constrained devices.
Design efficient models for microcontrollers.
Compress models through quantisation, pruning and distillation.
Deploy models with embedded frameworks and accelerators.
Manage device fleets and OTA updates.
Industrialise products with embedded AI securely.
Who it's for
- Edge AI engineers looking to deepen their expertise.
- TinyML developers moving from prototypes to production devices.
- Embedded systems engineers adding machine learning to their products.
- Embedded vision engineers.
- IoT architects.
- Connected product engineers.
- Engineering or science graduates aiming to enter consumer electronics.
What you'll work with
- Directive 2014/53/EU (radio equipment)
- Regulation (EU) 2024/2847 (Cyber Resilience Act)
- ETSI EN 303 645 (cyber security for consumer IoT)
- IEC 61508 (functional safety)
- PSA Certified (IoT security certification framework)
- Regulation (EU) 2024/1689 (AI Act)
- ISO/IEC 22989 (AI concepts and terminology)
- TensorFlow Lite for Microcontrollers
- Edge Impulse
- STM32Cube.AI
- Apache TVM
- NVIDIA TensorRT
12 modules
1.1Edge, fog and cloud1.2Latency, privacy and energy advantages1.3Device classes: MCU, MPU and accelerators1.4Memory and compute constraints1.5TinyML use cases1.6Typical toolchain1.7Energy consumption of inference1.8Ecosystem and communities1.9Limitations of ML on microcontrollers1.10Case study: platform selection for a smart sensor
2.1ARM Cortex-M architectures2.2DSP and Helium extensions2.3RISC-V microcontrollers2.4Embedded NPUs: Ethos-U2.5Arduino Nano 33 BLE and ESP32-S3 boards2.6Flash and SRAM memories2.7Digital sensors and interfaces2.8Power consumption measurement2.9Selection by cost and availability2.10Lab: power consumption characterisation of a board
3.1Audio, IMU and vision acquisition3.2Fixed-point preprocessing3.3MFCC and spectrogram extraction3.4Windowing and filtering3.5Field dataset collection3.6Efficient labelling3.7Data augmentation3.8Imbalance and unknown classes3.9Edge Impulse as an integrated workflow3.10Practical exercise: vibration dataset with an accelerometer
4.1Lightweight convolutional networks4.2MobileNet and depthwise separable convolutions4.3MCUNet architectures4.4Compact recurrent networks4.5Trees and classical models on MCUs4.6Neural architecture search4.7MAC and memory budget4.8Accuracy–latency trade-offs4.9MLPerf Tiny benchmarks4.10Lab: keyword spotting with a 20 kB CNN
5.1Post-training quantisation5.2Quantisation-aware training5.38-bit and 4-bit quantisation5.4Structured and unstructured pruning5.5Knowledge distillation5.6Weight factorisation5.7Graph optimisation5.8Mixed precision5.9Accuracy verification after compression5.10Practical exercise: int8 quantisation of a vision model
6.1TensorFlow Lite for Microcontrollers6.2LiteRT and model conversion6.3CMSIS-NN6.4STM32Cube.AI6.5ONNX and TVM compilers6.6Apache TVM microTVM6.7Memory and tensor arena management6.8RTOS integration6.9On-device debugging6.10Lab: deployment on STM32 with Cube.AI
7.1NVIDIA Jetson and TensorRT7.2Google Coral Edge TPU7.3Hailo and other NPUs7.4Embedded Linux and Yocto7.5Containers at the edge7.6Real-time computer vision7.7Multi-model inference7.8Thermal management7.9Performance-per-watt comparison7.10Practical exercise: object detector on Jetson Orin
8.1Keyword spotting8.2Predictive maintenance on sensors8.3Human activity recognition8.4Low-power vision8.5Precision agriculture8.6Environmental monitoring8.7Health wearables8.8Smart cities8.9Documented industrial cases8.10Case study: anomaly sensor for a motor
9.1BLE, LoRaWAN and NB-IoT9.2Sending inferences versus raw data9.3OTA model updates9.4Device fleet management9.5Drift monitoring at the edge9.6On-device learning9.7Federated learning on devices9.8Battery and energy harvesting9.9Product life cycle9.10Workshop: OTA update of a model in the field
10.1Secure boot and signed firmware10.2Model protection10.3Physical and side-channel attacks10.4PSA Certified and Arm TrustZone10.5Radio Equipment Directive10.6Product cyber resilience10.7Reliability in extreme conditions10.8Functional safety and ML10.9Data privacy at the edge10.10Case study: threat model for a TinyML node
11.1Hardware design with embedded AI11.2CE and FCC certification11.3Bill of materials cost11.4Production testing11.5Technical documentation11.6Manufacturing scale-up11.7Edge AI business models11.8Product project management11.9Sustainability and e-waste11.10Practical exercise: industrialisation plan for a device
12.1Definition of the device and use case12.2Hardware selection12.3Data capture12.4Model design and training12.5Optimisation and compression12.6Embedded deployment12.7Power consumption and latency measurements12.8Field validation12.9Thesis report12.10Master's Thesis: defence before a panel
How you'll learn
100% online training 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 each intake, so you can combine the programme with your job.
Each module combines theory, guided reading of the original standard and applied exercises. The working software is TensorFlow Lite for Microcontrollers, Edge Impulse, STM32Cube.AI, Apache TVM and NVIDIA TensorRT, in free, academic or evaluation versions. Testing is studied on real procedures for on-board power and latency measurements, field validation of models and MLPerf Tiny benchmarks. Assessment is continuous by module, plus the final Master's Thesis.
Standards covered: Directive 2014/53/EU, Regulation (EU) 2024/2847, ETSI EN 303 645, IEC 61508, PSA Certified, Regulation (EU) 2024/1689, ISO/IEC 22989.
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
- University degree in engineering, science or architecture, or equivalent demonstrable professional experience.
- Internet connection and a computer able to run 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 1,900 hours over 12–14 months. You set your own pace within each intake, and your tutor helps you plan the workload around your job.
You work with TensorFlow Lite for Microcontrollers, Edge Impulse, STM32Cube.AI, Apache TVM and NVIDIA TensorRT, 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 Master's Thesis. 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 in a course or postgraduate diploma in the same area are recognised in the higher programme. Ask about the recognition policy when you enrol.
A university degree in engineering, science or architecture (or equivalent demonstrable professional experience), a computer able to run the 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
Master’s in AI-Based Predictive Maintenance Engineering
Online master's programme of 1,900 hours in AI-based predictive maintenance: 12 modules covering maintenance data, feature engineering, fault diagnosis, asset anomaly detection, remaining useful life prognosis,…
Master’s in Multiphysics Co-Simulation & FMI/FMU
Online master's programme of 1,900 hours in multiphysics co-simulation and FMI/FMU: 1D/3D coupling, orchestration and HIL/SIL. 12 modules covering coupled multiphysics systems, Modelica modelling, the FMI…
Master’s in Factory and Supply Chain Digital Twins Engineering — Discrete-Event Simulation, WIP/OEE and What-If
Online master's programme of 1,900 hours in factory and supply chain digital twin engineering, with discrete-event simulation, WIP/OEE and what-if analysis. 12 modules covering discrete-event simulation,…
More in Data, AI, Simulation & Analytics
All Data, AI, Simulation & Analytics programmes → · How to apply · Fees & financing · Scholarships