Recommended prerequisite for participation in the module

Basics of Audio Signal Processing, Machine Learning, and AI Programming in python / MATLAB.

Prerequisite for participation in the module

AI Programming in python / MATLAB.

Content, progress and pedagogy of the module

The module is aimed at participants who are already in the workforce within audio software engineering, embedded systems, hearing-aid technology, and telecommunications. These professionals increasingly encounter complex workflows that involve multiple data types (e.g. sensor, audio, metadata), model orchestration, and multi-stage inference pipelines.

The module integrates concepts of workflow engineering and intelligent orchestration, applied in AI Engineering context to the need of modular, transparent, and automated model management pipelines in high-performance and high-reliability settings, for deploying and maintaining AI-based products and enterprise solutions.

Learning objectives

Knowledge

Upon completion of the module, participants will:

• have advanced theoretical knowledge of MLOps principles, focusing on audio-specific machine learning workflows
• understand the technical and organizational challenges in deploying AI/ML models within the audio industry
• prepare for real-world constraints like latency, reliability, and data drift
• gain a deep understanding of lifecycle management and ethical considerations in deploying machine learning in industry settings

Skills

Participants will be able to:

• develop and manage audio ML pipelines by incorporating best practices in model training, deployment, and monitoring
• perform robust feature extraction for audio datasets, using advanced tools such as TensorFlow, Kubernetes, and Docker, and evaluate the environmental impacts of these workflows 
• communicate complex ML engineering concepts and solutions effectively to both technical and non-technical stakeholders.

Competences

Participants will:

• demonstrate autonomy and responsibility in managing complex MLOps projects for audio-based systems, integrating knowledge from multiple disciplines (signal processing, machine learning, and systems design)
• develop collaborative skills by working within interdisciplinary teams and integrating input from audio engineers, AI researchers, and business managers 
• engage in continuous learning and reflection to adapt MLOps workflows to evolving audio industry trends, including sustainability and ethical AI concerns

Type of instruction

Lectures, classroom instruction, workshops, exercises (individually and in groups) and reflection.

Extent and expected workload

150 hours.

Exam

Prerequisite for enrollment for the exam

• None

Exams

Name of exam	Signal Processing for AI Engineering
Type of exam	Written exam
ECTS	5
Permitted aids	See semester description.
Assessment	Passed/Not Passed
Type of grading	Internal examination
Criteria of assessment	The criteria of assessment are stated in the Examination Policies and Procedures

Additional information

The module is a separate module and is placed at Master's level. There is a participant payment on the module.