KEY RESPONSIBILITIES
Generative AI Agentic Engineering (70%)
Design, prototype and build Generative AI applications, including prompt engineering and prompt optimization for reliability, accuracy and cost.
Build and tune Retrieval-Augmented Generation (RAG) pipelines, including chunking, embeddings, vector search and grounding strategies.
Design and build single- and multi-agent systems; implement agent-to-agent communication and orchestration patterns across tools, workflows and systems.
Implement memory management strategies for agents (short-term, long-term and episodic memory) to support context retention and coherent multi-turn behaviour.
Set up observability for AI/agentic systems: tracing, logging, evaluation and monitoring of prompts, model outputs, agent actions and failure modes.
Establish governance of agents, including permissioning, escalation paths, human-in-the-loop checkpoints and accountability for autonomous actions.
Implement guardrails for AI and agentic systems covering input/output validation, content safety, jailbreak and prompt-injection defence, and tool-use restrictions.
Apply software development practices to AI systems, including API integration, version control, CI/CD, testing and secure coding for AI/agentic applications.
AI Governance internal stakeholder alignment (30%)
Maintain the enterprise AI use case inventory / model registry; govern the AI solution lifecycle including periodic post-deployment monitoring, revalidation and decommissioning.
Ensure AI solutions comply with applicable regulations and guidelines (including IRDAI directions and emerging AI regulations) and internal risk appetite.
Identify, report and support response to AI-related incidents (model failures, misuse, unsafe agent behaviour) and drive corrective action; govern remediation and closure of AI-related risks and exceptions within timelines.
Drive responsible AI training, awareness and adoption across business functions.
