Role Overview
Capco is seeking an AI Engineer – Generative AI & Agentic Systems to support Sun Life’s Individual Insurance New Business transformation. You will design, build, test and deploy secure, production-ready AI and agentic capabilities that integrate with workflow orchestration, case management, decision services and enterprise systems in a regulated environment.
Responsibilities
- Build AI and agentic services using Python, approved cloud platforms, retrieval, structured outputs, tool calling, orchestration and state management.
- Design APIs, asynchronous services and integrations with workflow engines, case-management platforms, rules services, document processing and systems of record.
- Implement document classification, extraction, reconciliation and retrieval for insurance assessment and triage.
- Embed responsible AI controls: evaluation datasets, regression tests, human-review workflows, authorization and guardrails.
- Deliver production capabilities with CI/CD, monitoring, logging, tracing and operational documentation; test normal, exception and failure scenarios.
Requirements
- 3+ years in professional software, data or ML engineering, including 1+ year hands-on AI/ML and practical Generative AI, LLM or agentic delivery.
- 2+ years Python development and 2+ years building APIs, asynchronous services or enterprise integrations; FastAPI or equivalent, SQL, structured/unstructured data processing.
- 2+ years with AWS, Azure or GCP; Git, Docker, CI/CD, automated testing, secure coding, secrets management and production troubleshooting.
- Applied AI knowledge: LLM APIs, structured outputs, tool/function calling, RAG, embeddings, vector retrieval, agent orchestration, state/memory patterns, prompt/configuration versioning and AI evaluation.
- Strong problem-solving, communication and documentation; ability to work iteratively with stakeholders and understand privacy, security, auditability, workflow decisioning and human oversight.
Bonus Points
- Amazon Bedrock, Bedrock AgentCore, LangGraph, Google ADK, Semantic Kernel or comparable managed AI/agent orchestration frameworks.
- Model Context Protocol (MCP), Neo4j, graph databases, agent-tool interoperability or evaluation of agent outputs and execution trajectories.
- Document intelligence, workflow and case-management platforms, event-driven architectures or durable messaging.
- Legacy-system integration, mainframe environments, infrastructure as code or enterprise modernization programmes.
- Financial services, regulated environments, life-insurance New Business, underwriting or client-facing engineering/consulting delivery.