• Design, build, own and operate production SaaS features that integrate RAG pipelines, agent systems including multi-agent architectures, and deterministic services. • Use AI coding agents (e.g., Claude Code style tools) as a first class part of your development process, including architecture exploration, implementation, testing, refactoring, and documentation, while maintaining full ownership of output quality and understanding their strengths and limitations. • Implement and evolve agentic systems using frameworks such as LangChain, CrewAI, PydanticAI, or equivalent, including tool orchestration, memory/state handling, and multi-step reasoning workflows. • Build scalable retrieval and inference pipelines: embedding workflows, vector databases, hybrid retrieval strategies, caching, and evaluation loops to balance accuracy, latency, and cost. • Identify where agentic approaches provide leverage over conventional code and where they do not, making pragmatic, production minded tradeoffs across cost, accuracy, latency, and reliability. • Design guardrails and failure-mode mitigations for non-deterministic systems, including prompt structure, validation layers, fallback strategies, and human-in-the-loop patterns where appropriate. • Integrate observability and evaluation into AI-enabled features, monitoring correctness, drift, bias, and performance over time. • Collaborate with architects, product managers, and designers to translate ambiguous requirements into robust technical designs. • Contribute to shared libraries, internal tooling, and reference patterns that accelerate delivery of AI-enabled features across teams. • Participate in code reviews and design discussions, raising the bar for quality, reliability, and responsible use of AI across the engineering organization. • Mentor other engineers on effective agentic coding practices, prompt design, and emerging patterns as adoption scales. • Own AI-enabled features end-to-end in production, from initial design and implementation through deployment, monitoring, on-call support, incident response, and iterative improvement.