Skip to main content

Skills application demo

Instruction and application
In Progress

Skills application demo​

Security Walkthrough​

In this skills application demo, we’ll explore the end-to-end threat model for LogiFleet’s ML system, which predicts delivery times using geolocation, traffic, and package data. The walkthrough highlights vulnerabilities across data ingestion, training, and deployment—then maps each risk to core security principles and practical mitigations. This activity reinforces how threat modeling translates into real-world decisions that strengthen ML system integrity, confidentiality, and resilience.

The walkthrough guide​

Select the link below to download the Security Walkthrough guide. It outlines the ML system architecture, key vulnerabilities, and their links to core security principles.

The guide also includes discussion questions and mitigation strategies to support the demo and group reflection.

Use the guide to follow along as the coach walks through the system and highlights key security issues.

Optional: Download a copy of the workshop slides​

Activity instructions​

Watch the walkthrough​

Watch the coach demonstrate how to identify vulnerabilities in the system, map them to core security principles, and recommend effective mitigation strategies.

Participate in the discussion​

Consider the following as you reflect and contribute:

  • Where do you see vulnerabilities in this system?
  • Which security principles are at risk at each stage?
  • What mitigation strategies would you recommend?
  • What new risks might emerge if the system expands globally?

Key takeaways​

  • **Routine components can introduce major risks:**Weak API security, unencrypted streams, and open access points often expose ML systems to attack.
  • **Vulnerabilities map to core security principles:**Each issue reflects a breach in confidentiality, integrity, authentication, non-repudiation, or service integrity.
  • **Layered mitigation is essential:**Combining encryption, access control, input validation, and logging builds resilient ML workflows.

Skills application​

Securing an ML system​

In this skills application, you’ll analyse an ML system used for route optimisation, identifying security vulnerabilities across the ML lifecycle. You’ll map each vulnerability to a core security principle and propose appropriate mitigation strategies, both technical and procedural.

Learning this skill supports more resilient ML practices by helping teams anticipate risks, design safeguards, and align system behavior with organisational security goals.

The Challenge guide​

Select the link below to download and access the Securing an ML system challenge guide. It includes the challenge instructions and the content needed to complete the tasks including the system overview, security gaps, and structured prompts to guide your analysis and design.

Optional: Download a copy of the workshop slides​

Activity instructions​

Work on the challenge tasks​

With your group, use the challenge guide to review the PackSure ML system. Identify three key vulnerabilities, link each to a core security principle, suggest one technical and one process-based mitigation, and outline a basic team protocol for monitoring and response.

Collaborate in the breakout room​

Discuss risks, compare your thinking, and co-develop practical mitigation strategies for PackSure’s ML workflow. Help each other align recommendations with real-world constraints.

Share your findings​

Submit your completed PackSure Secure ML Plan. It should capture the key risks, mapped principles, mitigation strategies, and team protocol.

Regroup​

Return to the main session after 20 minutes to discuss key takeaways and insights from the challenge.