AI safety & Security controls, LLM Guardrails/Associate Director, Software Engineering Specialist

Some careers shine brighter than others.

If you’re looking for a career that will help you stand out, join HSBC and fulfil your potential. Whether you want a career that could take you to the top, or simply take you in an exciting new direction, HSBC offers opportunities, support and rewards that will take you further.

HSBC is one of the largest banking and financial services organisations in the world, with operations in 64 countries and territories. We aim to be where the growth is, enabling businesses to thrive and economies to prosper, and, ultimately, helping people to fulfil their hopes and realise their ambitions.

We are currently seeking an experienced professional to join our team in the role of Associate Director, Software Engineering Specialist

In This Role, You Will

• Design, implement and operate AI safety detectors for enterprise AI use cases, including prompt injection and jailbreak detection, personally identifiable information and sensitive data detection, data leakage prevention, content safety and harmful content detection, policy violation checks, system prompt leakage detection, and agent or tool-use risk controls.
• Implement detector pipelines using a combination of rules, machine learning models, LLM-based evaluators, and third-party safety services, with clear interfaces, confidence scoring and policy-driven actions.
• Deliver evaluation workflows for guardrails, including offline test sets, online validation, threshold tuning, regression checks, and false positive and false negative analysis to measure detector effectiveness in production.
• Design end-to-end protection patterns for enterprise retrieval-augmented generation, coding assistants and agent systems across upload, retrieval, prompt, response, logging and tool execution layers.
• Build policy enforcement and decisioning services that support configurable actions such as allow, block, redact, mask, escalate for review or require human approval.
• Build observability and governance capabilities for AI safety, including telemetry, tracing, dashboards, audit trails, review workflows, red team feedback loops and operational controls.
• Partner with platform, infrastructure, security and AI teams to productionise guardrails with performance controls such as parallel detector execution, tiered detection, graceful degradation, timeout isolation and scalable rollout strategies.

To be successful in this role, you should meet the following requirements:

• Bachelor’s degree or above in Computer Science, Software Engineering, Artificial Intelligence, Machine Learning or a related discipline, or equivalent practical experience with strong software engineering experience building and operating distributed, cloud-native backend systems in production.
• Strong hands-on experience building and operating production AI safety, guardrail, content safety, trust and safety, or AI risk detection systems, rather than only providing architectural guidance, high-level solution design, or developing general-purpose AI applications.
• Practical experience in several of the following areas: prompt injection detection, jailbreak defence, personally identifiable information detection, sensitive data or data leakage prevention, content safety or harmful content detection, policy violation or response validation, system prompt leakage detection, or agent safety controls.
• Experience implementing detection approaches such as rule engines, regular expressions or pattern matching, named entity recognition, classifiers, LLM-as-judge patterns, or hybrid detector ensembles, and understanding where each approach is effective and where it breaks down.
• Strong understanding of evaluation and measurement for safety systems, including precision, recall, threshold design, false positive and false negative trade-offs, offline and online evaluation, and human review workflows. Experience building or supporting safety controls for retrieval-augmented generation, enterprise search, coding assistants, copilots or agent-based systems.
• Demonstrated direct ownership of implementation and productionisation, including developing or integrating detectors, tuning thresholds, building evaluation workflows, troubleshooting production issues, and supporting rollout and operational governance.
• Candidates with additional experience in developing first-party detectors, building fine-tuning pipelines for safety or detector use cases, or evaluating an…