This role is best suited to someone who combines strong engineering judgment with real practical experience in security, governance, privacy, or compliance for AI-powered systems. We care more about evidence of sound decision-making and pragmatic implementation than about titles or years in a specific niche.

What youʼll do

• Define and improve practical guardrails and safe defaults for AI-powered systems
• Help shape Awinʼs approach to privacy, PII handling, and data-safe patterns for AI workflows
• Contribute to approved provider and model guidance, including practical validation and usage expectations
• Work with engineers to translate policy, risk, and governance needs into reusable platform patterns and engineering controls
• Help define production-ready expectations for AI systems from a safety, governance, and risk perspective
• Contribute to readiness checks, automated validations, and scalable governance mechanisms that reduce reliance on manual approval processes
• Help improve logging, tracing, and observability practices so AI systems are auditable and safer to operate
• Partner closely with Security, Legal, Architecture, and product engineering teams to ensure standards are practical and proportionate
• Support teams in understanding AI-specific risks such as unsafe outputs, provider misuse, privacy exposure, and poor trace hygiene
• Write clear documentation, patterns, and guidance that help teams apply standards consistently
• Contribute to strong engineering practices across the team through clean code, testing, and collaboration 
• Mentor others and help raise the teamʼs overall maturity in AI safety, governance, and privacy-aware engineering

Requirements

• Strong software engineering fundamentals and experience working on production systems
• Practical experience in security, governance, privacy, compliance, or risk-related engineering work
• Clear hands-on experience applying those controls to AI- or LLM-powered systems
• Strong understanding of AI-specific risks, including privacy exposure, PII handling, unsafe outputs, provider/model risk, and misuse or abuse patterns
• Experience translating policy or risk requirements into practical engineering controls, patterns, or defaults
• Good understanding of how to make AI systems observable, auditable, and safer to operate in production
• Ability to work effectively with engineers, architects, and non-engineering stakeholders such as Security or Legal
• Good practical experience with Python and/or JavaScript / TypeScript
• Comfortable working across the lifecycle of a platform capability, from design and implementation to rollout and iteration
• Good understanding of information security and how to design solutions with security in mind
• Comfortable applying unit testing, continuous integration, and continuous deploymentStrong communication skills, both synchronous and asynchronous
• Comfortable working through ambiguity and helping shape standards where the right balance between safety and speed is still evolving
• Strong judgment and pragmatism, with an ability to improve safety without introducing excessive bureaucracy

Nice to have

• Experience designing or implementing guardrails for AI or LLM-powered systems
• Experience with AI governance automation, readiness gates, policy checks, or preflight validation patterns
• Familiarity with tracing, auditability, and logging standards for AI systems
• Familiarity with evaluation workflows, regression checks, and production readiness practices for AI
• Experience assessing third-party AI providers or model usage from a risk or governance perspective
•  Experience with privacy-by-design patterns in data-sensitive systems
• Experience working in regulated or higher-risk environments
• Experience writing clear developer guidance and turning repeated risk questions into reusable standards

Tools and technologies you may work with:

• Python
• JavaScript / TypeScript
• Azure / AWS
• LLM APIs
• CI/CD tooling
• Observability and tracing tools
• Policy and validation patterns
• Internal platform services, standards, and governance tooling
• Working style