SynapsedDecide and govern
Managers, Chief AI Officers, product owners and governance teams.
Synapsed · Trustworthy AI Training
Build the people behind trustworthy AI.
A capable AI team is the foundation of a mature AI initiative. Give leaders, engineers and assessors the knowledge and practical skills to make AI responsible, reliable and governable.
Instructor ledRole-based learningPractical exercises
Three learning paths. One connected AI team.
Start with the responsibilities your people hold. Take a focused course or combine programmes into a learning path for teams working together.
Design and build
AI engineers, architects, developers and analysts.
Evaluate and improve
AI auditors, developers, risk and quality professionals.
Trustworthy AI Masterclass
Connect governance to the AI lifecycle.
For people leading AI initiatives, the masterclass brings together governance, risk, data quality, lifecycle methods and threat modeling. Work through examples of design failures and explore how to assess organisational readiness.
The Synapsed governance approach helps connect these questions across five dimensions:
- AI Team
- Processes
- Guidelines
- Standards
- Guardrails
From informed adoption to practical engineering.
Explore each programme’s focus, audience and starting knowledge. The level of detail and examples can be tailored to your organisation.
Introduction to AI: Foundations and Strategies for Successful Adoption
Build a shared understanding of AI, its impact on the business and the decisions behind responsible adoption. Connect opportunities with the capabilities and governance needed to pursue them.
Explore the programme and learning outcomes
Programme focus
- Traditional software and AI: differences, business impact and implications for users.
- Organisational AI risks and approaches to risk and impact assessment.
- An introduction to the EU AI Act, ISO/IEC 42001, ISO/IEC 5338 and NIST AI RMF.
- Governance methods, transparency, explainability, fairness and privacy by design.
Learning outcomes
Recognise the questions an organisation should address before adopting AI and contribute to an informed discussion of risks, responsibilities and priorities.
Starting knowledge
Basic familiarity with the software development lifecycle is recommended.
AI Governance Framework: Develop Trustworthy AI Systems
Connect AI ambitions with accountability, oversight and practical governance. Explore how an organisation can assess its readiness and translate trustworthiness objectives into a working framework.
Explore the programme and learning outcomes
Programme focus
- AI risks: bias, transparency, accountability and privacy, considered alongside business objectives.
- The relationship between governance, regulation, standards and the AI lifecycle.
- Roles, responsibilities, policies and guidelines for responsible AI.
- Synapsed’s governance approach and organisational maturity assessment, including OWASP AIMA.
Learning outcomes
Outline a governance framework, identify ownership and review mechanisms, and use a maturity perspective to prioritise improvements.
Starting knowledge
No specific technical background is required. Familiarity with AI applications and business processes is useful.
Build Trustworthy AI Systems
Bring trustworthiness into engineering decisions throughout the AI lifecycle. Connect data preparation, model behaviour and development practices with the needs of users and stakeholders.
Explore the programme and learning outcomes
Programme focus
- Trustworthiness principles and their integration from design through deployment.
- Data quality and integrity, privacy, preparation and feature engineering.
- Bias detection and mitigation, model interpretability and transparent communication.
- Robustness, resilience, threat modeling and development practices informed by relevant requirements.
Learning outcomes
Identify practical development measures for reliable, fair and understandable AI, and explain how these measures fit into the wider lifecycle.
Starting knowledge
Machine learning concepts, programming experience and basic AI system design. Familiarity with data handling and software development is recommended.
Design Trustworthy AI Systems: AI Threat Modeling
Examine how an AI system could behave outside its intended boundaries. Use structured analysis and group exercises to connect threats, risks and design decisions.
Explore the programme and learning outcomes
Programme focus
- Threat modeling concepts and techniques, including STRIDE and attack trees.
- Threats affecting data, models and LLM applications, including poisoning and prompt injection.
- Assessing likelihood and impact, documenting assumptions and prioritising mitigations.
- A group exercise to identify and document threats in a sample AI system, then discuss controls and validation.
Learning outcomes
Produce a structured threat model for an example AI system and connect prioritised risks with mitigations, design choices and subsequent tests.
Starting knowledge
Basic AI and risk concepts. Familiarity with AI system design and the software development lifecycle is recommended.
05 / Practical assessmentTrustworthy AI:
Trustworthy AI:
Testing, Evaluation & Assurance
Learn to examine the behaviour of AI applications, test their limits and evaluate the controls that support trustworthy operation. Connect structured testing with evidence, improvement and lifecycle decisions.
Understand and evaluate
- AI threat modeling and risk scenarios for LLM applications and agents.
- Prompt injection, data and supply-chain weaknesses, and unintended outputs.
- OWASP AI Testing Guide methods and examples from the OWASP Top 10 for LLMs.
- Hands-on testing of a purpose-built educational AI application.
Control and improve
- Architectural controls for LLMs and agents: layers, proxies and sandboxes.
- Guardrails, input and output filtering, policies and permissions.
- Control implementation and verification across the AI lifecycle.
- Compare behaviour before and after a change; discuss evidence and remaining limitations.
Learn through the Synapsed AI Lab.
A dedicated learning environment connects concepts with observed behaviour. Participants work through practical scenarios, inspect results and discuss how to improve the application.
- Understand: examine the application and select a relevant scenario.
- Evaluate: run a test and inspect the system’s response.
- Improve: introduce a control linked to the observed weakness.
- Verify: repeat the scenario, compare the evidence and discuss limitations.
Explore the learning sequence and prerequisites
- Foundations
- Differences between traditional software and AI; trustworthiness objectives; threat modeling; LLM and agent risk scenarios.
- Guided assessment
- OWASP AITG methodology and testing tools, applied to a dedicated training application with practical exercises.
- Controls and review
- Lifecycle controls, guardrail implementation and review of results from the lab.
A general understanding of AI and familiarity with assessment or auditing processes is helpful. Duration, delivery format and technical requirements are agreed for the programme. Individual lab access is provided for a limited period, with access dates confirmed for each course.
Keep learning after the session.
Participants receive the presentation materials. Technical courses include the exercises discussed during training; a certificate of participation is available on request for the training courses.
The introductory workshop includes its presentation materials. Lab access applies to the practical testing programme and is time limited.
Connect understanding, practice and application.
Start from responsibilities
Align the audience, starting knowledge and learning objectives. Leaders, engineers and assessors need different depth and a shared vocabulary.
Work through examples
Use discussions, design scenarios and practical exercises to connect concepts with real decisions. The lab programme adds direct experience of testing and evaluating controls.
Apply it to your context
Relate the learning to your organisation’s AI initiatives. Identify the decisions, practices and follow-up work needed to put new skills into use.
Research and community
Learn from the work behind the methods.
Synapsed brings applied research, engineering methods and assessment experience into its training. Our work connects AI-T-SDLC, governance frameworks and evidence-based testing.
Matteo Meucci is Creator and Contributor of the OWASP AI Maturity Assessment and OWASP AI Testing Guide projects.
These are open OWASP community projects. The programmes presented here are Synapsed training offerings.
Choose the depth your team needs.
Where should a mixed team start?
The introductory workshop establishes shared language for AI adoption. Leaders can then focus on governance, while technical roles develop engineering or assessment skills. The masterclass connects these perspectives for people leading AI initiatives.
Do all courses require programming?
No. The governance course does not require a specific technical background. Build Trustworthy AI Systems expects programming experience and familiarity with machine learning. Each course lists its recommended starting knowledge.
Can the content be adapted to our organisation?
Yes. Agree the audience, objectives, examples and technical depth when planning the programme. Delivery arrangements and lab requirements are confirmed for the engagement.
How does training connect with AI Services?
Training develops understanding and skills. Assessment, engineering and assurance engagements apply related methods to your organisation or AI application. We can help identify whether your next step is a course, a practical review or a combination.
Develop the skills your AI initiative needs.
Tell us who needs to learn, what they are working on and the capabilities you want to develop. We will help identify a suitable learning path.