Sr Software Engineer

The Senior Software Engineer (IC3) on AI-Native ITSM designs, builds, and delivers agentic and conversational features that help IT operators and end-users navigate the ITSM workflows. You work within an established architecture and domain strategy to author agentic instructions, build conversational experiences, and integrate AI capabilities into production ITSM workflows. Unlike traditional full-stack roles, you engage directly with natural-language engineering—prompts, instructions, and tool definitions are your working materials, versioned and tested alongside code. You design and build the automated evaluation that verifies agentic behavior at scale, and you own the correctness and safety of the features you ship. You collaborate with IC4 architects on domain strategy and mentor IC1-IC2 engineers on the practices that make agentic delivery reliable. #### What You Do - Build agentic features within the incident and request lifecycle: Deliver complete features that span intent interpretation, multi-step reasoning, tool invocation, and action—such as intelligent request categorization, incident summarization and classification, or guided self-service resolution—working within the domain architecture established by IC4. - Author and test agentic instructions: Write system instructions, tool definitions, guardrails, and escalation rules as engineered artifacts. Design evaluation sets and automated testing that verify behavior across model versions and edge cases. Collaborate with IC4 on shared instruction surfaces. - Build conversational experiences: Design and implement experiences that hold context across turns, hand off cleanly to live agents, and behave consistently across chat, voice, and API channels. Account for user error recovery, confirmation before consequential actions, and graceful degradation. - Integrate AI capabilities into production ITSM workflows: Connect LLM APIs, retrieval-grounded features, agent orchestration, and tool calling to ITSM data models and business logic. Implement structured output enforcement and validation. - Design and operate automated evaluation: Build golden datasets, multi-turn test suites, and scoring frameworks calibrated to human judgment. Implement CI gates and drift detection that make it safe to iterate on instructions and models. - Participate in production operations: Monitor conversation quality, containment, and error rates. Investigate production issues, trace agentic reasoning, and feed findings back into specifications and evaluation sets. Participate in on-call rotations for customer-facing systems. - Mentor junior engineers: Guide IC1-IC2 engineers on agentic instruction authoring, testing practices, and responsible AI principles. Review code and specifications for correctness, security, and maintainability. - Contribute to platform architecture: Work with IC4 on data models, API design, and integration patterns that support agentic behavior at scale. Identify improvements and technical debt.

Required Experience and Skills - Production software delivery: 5–7 years building and shipping production features, including direct experience with AI-native application features used by real customers. - Agentic experience: Direct experience authoring prompts and instructions, designing autonomous workflows, or building evaluation for model-driven features. This is required, not preferred. - LLM integration: Experience with LLM APIs, agent orchestration, function calling, and retrieval augmentation in production systems. - Full-stack fundamentals: Strong proficiency in front-end development (modern frameworks), server-side development, data modeling, and API design (REST or GraphQL). - Applied ML literacy: Working understanding of evaluation metrics, embeddings, model behavior, and common failure modes sufficient to reason about and verify agentic behavior. - Testing and quality: Demonstrated ability to design comprehensive test suites, including automated testing of non-determin