Software Engineer II AI & Agentic Systems

Must-have:

• 3+ years building enterprise-grade software commercially, with a solid grasp of testing and QA.
• 3+ years of commercial experience in Java or Python.
• Hands-on experience building with LLM APIs and agentic frameworks – tool calling, multi-step execution, structured output – well beyond using AI coding assistants.
• Hands-on experience with AI-driven coding assistants (GitHub Copilot, Cursor, Claude Code) to speed up your own work.
• Sound judgment about where automation should and shouldn't act on its own – and the ability to articulate exactly where that line sits.
• You think in trust boundaries, blast radius, and least privilege by default.
• You treat vulnerability management as a discipline: severity and exploitability, SLAs, and exception handling.
• Experience in an agile organization and with build systems like Jenkins or Bamboo.
• You build trust and rapport across all levels, and do your best work collaboratively in a fast-paced environment.

Nice-to-have:

• Software supply chain security – SBOMs, provenance and attestation, image signing and verification, policy-as-code.
• Experience running AI coding agents at repository scale, and building the harness that makes their output trustworthy.
• Familiarity with AI governance in a security context – model-use policy, data handling, and the audit questions that follow an agent-initiated production change.
• Experience in a regulated environment (SOC 2, FedRAMP, or similar) and a feel for what auditors actually ask for.

What you’ll do at Jamf:

At Jamf, we empower people to be their best selves and do their best work. As a Software Engineer II on a new team enabling secure, governed adoption of autonomous and agentic workflows, you'll help build production systems where AI agents do real work on real code - authoring changes inside deterministic gates, with independent verification and a full audit trail. This is a greenfield build inside a mature platform, starting with Security.

,[Build the production systems where AI agents do real work on real code – authoring changes that run inside deterministic gates, with independent verification and a full audit trail behind every one. , Take customer and platform problems, break them into workable pieces with your agile team, and own most of them end to end. , Draw the line on autonomy: design the trust boundaries, blast radius, and least-privilege model for the automation itself, which holds elevated access to production. , Build the evaluation and observability that keeps non-deterministic systems honest – measuring accuracy and catching silent degradation before anyone else does. , Make the systems answer the auditor's "show me" – producing evidence that a control worked, not just the claim that it did. , Work shoulder to shoulder with security engineering and PDD developers, and explain the tricky parts clearly to the rest of the organization. , Bring engineering best practices to the table: ask the sharp questions, solve the hard problems, and help others level up. ] Requirements: Java, Python, Testing, Agentic frameworks, Github Copilot, Cursor, Claude Code, Jenkins, Bamboo, SBOM, SOC2, FedRAMP Tools: Agile. Additionally: International projects, Small teams, Apple equipment, Training budget, Private healthcare, Flat structure, Free coffee, Bike parking, Playroom, Shower, Free snacks, Modern office, No dress code, Kuchnia.