· Strong understanding of agentic AI fundamentals—goal definition, planning, reasoning, task decomposition, and autonomous action loops.
· Design and build agentic workflows using LLMs, including tool/function calling, state management, memory, and orchestration patterns (single\-agent/multi\-agent).
· Hands\-on experience with agent frameworks and patterns (e.g., LangChain\-based agent frameworks or equivalents), including routing, workflows, and agent coordination.
· Model Context Protocol (MCP) proficiency for integrating tools and contextual data into agent workflows (tool discovery/usage via MCP servers).
· Build Retrieval\-Augmented Generation (RAG) pipelines and gr ounding strategies for enterprise knowledge sources (e.g., internal repos/wikis/ticketing/content stores) to improve agent reliability.
· Agent observability \& monitoring (AgentOps): track agent behavior, tool calls, outcomes, and quality signals; implement alerts/traces and feedback loops for continuous improvement.
· Cloud and/or platform engineering exposure for deploying agentic systems at scale (containers/Kubernetes/OpenShift, CI/CD, performance tuning, and secure operations).
· Responsible AI \& guardrails: design human\-in\-the\-loop controls, safe tool access, and security\-aware policies for agents operating with credentials, data access, and budgets.
· Proficiency in programming languages such as Python, .Net or Java, with experience in relevant libraries and frameworks (e.g., TensorFlow, PyTorch, Keras).