Master Thesis – Performance Validation of Edge AI Video Analytics

Camera-based Edge AI can detect objects and generate events in real time, but its performance varies with environmental conditions, installation and system load. Unlike model development, live deployments typically lack annotated datasets for evaluation. This thesis aims to create an automated validation framework that uses live video and controlled tests to assess end-to-end analytics performance and its relevance to real-world deployments.

The assignment

• Investigate existing approaches to AI benchmarking and validation.
• Design a standardized methodology for testing complete Edge AI video analytics systems.
• Build a prototype that automates test execution, data collection and performance analysis.
• Use statistical techniques to examine variations in analytics performance.
• Document findings and recommend directions for future validation frameworks.

Who we are looking for

• Master’s students in computer science, robotics and automation, data science, applied mathematics or a related field.
• Strong analytical abilities and an interest in AI-based systems.
• Experience or interest in computer vision, machine learning, deep learning, statistics, data analysis or software engineering.
• Enrolled at a Swedish university or university college, either directly or through an exchange programme.

What we offer

• A thesis project combining computer vision and machine learning with software development and applied statistics.
• Academic project supervision with a named supervisor at Axis.
• An open, friendly and collaborative company culture that encourages new ideas and continued learning.

Practical details

Location: Lund. For one or two students, each completing a 30-credit master’s thesis. Apply through the thesis advert by 27 November 2026; applications may be written in Swedish or English. Include your CV and university transcript. If applying as a pair, submit separate applications and identify your thesis partner in each. Questions: supervisor Giuseppe Bianco (giuseppe.bianco@axis.com) or manager Carl Wiberg (carl.wiberg@axis.com).