Postgraduate Diploma in TinyML: Quantisation, Pruning and On-Device Compilation
Online 800-hour Postgraduate Diploma in tinyML: Quantisation, Pruning and On-Device Compilation. 8 modules covering Regulation (EU) 2024/1689 (AI Act), Regulation (EU) 2024/2847 (Cyber Resilience Act) and TensorFlow Lite for Microcontrollers. Dedicated tutor and open enrolment.
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
Explain the memory, compute and energy constraints of embedded machine learning.
Select efficient model architectures for microcontrollers.
Apply post-training quantisation and quantisation-aware training.
Reduce model size through pruning, distillation and compression.
Compile and deploy models with TensorFlow Lite Micro, TVM and vendor tools.
Measure system performance in terms of latency, memory, accuracy and power consumption.
Develop a complete TinyML system from data to device.
Who it's for
- TinyML Engineer seeking to specialise in TinyML: Quantisation, Pruning and On-Device Compilation.
- Embedded Systems Engineer seeking to specialise in TinyML: Quantisation, Pruning and On-Device Compilation.
- Edge Machine Learning Engineer seeking to specialise in TinyML: Quantisation, Pruning and On-Device Compilation.
- IoT Firmware Developer seeking to specialise in TinyML: Quantisation, Pruning and On-Device Compilation.
- Model Optimisation Engineer seeking to specialise in TinyML: Quantisation, Pruning and On-Device Compilation.
- Smart Device Architect seeking to specialise in TinyML: Quantisation, Pruning and On-Device Compilation.
- Graduates in engineering or science seeking to enter the consumer electronics sector.
What you'll work with
- Regulation (EU) 2024/1689 (AI Act)
- Regulation (EU) 2024/2847 (Cyber Resilience Act)
- ISO/IEC 22989 (concepts and terminology of AI)
- ISO/IEC 23053 (framework for AI systems using machine learning)
- IEC 61508 (functional safety)
- ETSI EN 303 645 (IoT cybersecurity)
- MLPerf Tiny (reference benchmark)
8 modules · 800 h
1.1From cloud to edge and to the microcontroller1.2Memory, compute and energy constraints1.3Cortex-M and RISC-V microcontrollers1.4Embedded NPU accelerators1.5Use cases: audio, vision and vibration1.6TinyML development cycle1.7Frameworks: LiteRT Micro and Edge Impulse1.8Latency and power consumption metrics1.9MLPerf Tiny benchmark1.10Practical case: keyword spotting on a microcontroller
2.1Lightweight convolutional networks2.2MobileNet and depthwise separable convolutions2.3Microcontroller architectures: MCUNet2.4Constrained architecture search2.5Models for time series2.6MFCC feature extraction2.7Knowledge distillation2.8Anomaly models2.9Accuracy–size trade-off2.10Practice: training a compact model in Keras
3.1Fixed-point representation3.2Post-training quantisation3.3Symmetric and asymmetric int8 quantisation3.4Per-channel quantisation3.5Quantisation-aware training3.64-bit and binary quantisation3.7Calibration with representative data3.8Accuracy loss and debugging3.9Supported operators3.10Laboratory: int8 quantisation of a model
4.1Unstructured pruning4.2Channel-structured pruning4.3Progressive magnitude pruning4.4Lottery ticket hypothesis4.5Weight clustering4.6Compression and encoding4.7Sparsity and real acceleration4.8Combining techniques4.9TensorFlow Model Optimization toolkit4.10Practice: structured pruning and latency measurement
5.1Interpreters versus compilers5.2Apache TVM and microTVM5.3CMSIS-NN and optimised kernels5.4Memory management and tensor arena5.5Operator fusion5.6STM32Cube.AI tools5.7Compilation for embedded NPUs5.8On-board profiling5.9RTOS integration5.10Laboratory: deployment on a Cortex-M board
6.1Data acquisition from sensors6.2On-device pre-processing6.3Energy management and duty cycles6.4Event-driven inference6.5BLE and LoRa communication6.6Model updates6.7On-device learning6.8Basic model security6.9Testing and validation6.10Practice: vibration anomaly detector
7.1Predictive maintenance7.2Health wearables7.3Smart agriculture7.4Low-power vision7.5MLOps for the edge7.6Drift monitoring7.7Sustainability and footprint7.8Certification and regulation7.9Product costs7.10Practical case: commercial TinyML product
8.1Definition of the use case8.2Data capture8.3Model design8.4Optimisation and quantisation8.5Compilation8.6Board deployment8.7Latency and power consumption measurement8.8Analysis of results8.9Report and presentation8.10Final review with the tutor
How you'll learn
Training 100 % online on the SEIUM campus, with downloadable material, recorded classes and a dedicated technical tutor. No fixed timetable: students progress at their own pace within the intake, which allows it to be combined with work.
Each module combines theory, guided reading of the original standard and applied exercises. The working software is TensorFlow Lite for Microcontrollers, Edge Impulse, Apache TVM, STM32Cube.AI, PyTorch, in free, academic or evaluation versions. Tests are studied through real procedures for deploying models on actual microcontrollers, measurement of latency, memory and energy consumption, and comparison of quantised and pruned models. Assessment is continuous per module plus the final integrating project.
Build it end to end
Your final project follows the real engineering workflow, step by step, reviewed by your tutor.
Explain the memory, compute and energy constraints of embedded machine learning.
Select efficient model architectures for microcontrollers.
Apply post-training quantisation and quantisation-aware training.
Reduce model size through pruning, distillation and compression.
Compile and deploy models with TensorFlow Lite Micro, TVM and vendor tools.
Where this takes you
Requirements
- Prior technical training or professional experience in the sector. No university degree is required.
- Internet connection and a computer capable of running the programme's software.
- Sufficient level of Spanish to follow technical regulatory reading.
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 spread over 6 to 8 months. That is an average of 10 to 15 hours per week, compatible with a full-time job.
The course uses TensorFlow Lite for Microcontrollers, Edge Impulse, Apache TVM, STM32Cube.AI and PyTorch. Free, academic or evaluation versions are used, and the tutor indicates the free alternative where one exists.
An own diploma from SEIUM upon passing the continuous assessment and the integrating project. It is in-house training, not an official qualification from the Ministry.
Yes. Modules passed in a course or postgraduate diploma in the same area can be recognised towards the advanced programme. Check the recognition policy when you enrol.
Related programmes
Postgraduate Diploma in Logistics, Kits and Patient Traceability
Online 800-hour postgraduate diploma in Logistics, Kits and Patient Traceability. 8 modules covering the Guidelines on Good Distribution Practice of Medicinal Products (2013/C 343/01), EudraLex Volume…
Postgraduate Diploma in Thermal/Ergonomic Comfort and Textile Regulations
Online 800-hour Postgraduate Diploma in Thermal/Ergonomic Comfort and Textile Regulations. 8 modules covering ISO 11092 (thermal and water vapour resistance, sweating hot plate), ISO 15831 (thermal…
Postgraduate Diploma in Accessibility Audits and Continuous Improvement
Online 800-hour Postgraduate Diploma in Accessibility Audits and Continuous Improvement. 8 modules covering UNE-EN 301549 (accessibility requirements for ICT products and services), WCAG 2.2 and axe…
More in Health, Biomechanics, Wearables & Operational Safety
All Health, Biomechanics, Wearables & Operational Safety programmes → · How to apply · Fees & financing · Scholarships