Senior Agentic AI Engineer
hsbc
Job Description
- Design and implement production-grade multi-step agent workflows, including planner and executor patterns and multi-agent coordination.
- Build secure tool-calling integration layers so agents can diagnose incidents and execute controlled remediation actions.
- Implement RAG pipelines that ground agent decisions with topology, runbooks, incident history, and change context.
- Define and maintain prompt templates, structured output contracts, and evaluation harnesses for regression-safe releases.
- Integrate LLM models based on latency, cost, and risk constraints, and optimise inference behaviour for operational use.
- Implement AI safety controls including approval gates, least privilege access, deterministic guardrails, and blast radius limits.
- Integrate agent workflows with observability platforms, ITSM tools, and CI/CD pipelines for end-to-end AIOps execution.
- Build complete auditability for all agent decisions, tool calls, approvals, actions, and rollback outcomes.