AI Tools and Methodologies Engineer

ZON builds AI technologies and products, including software that optimizes data center energy use. The AI Tools and Methodologies Engineer evaluates modern AI tools and methods, builds reusable tooling and evaluation approaches, and drives their adoption across the company.

Responsibilities:

• Track, evaluate and pilot coding agents, agent frameworks, protocols and developer tools, and recommend what we adopt. Examples of the type of tool: agent orchestration frameworks, MCP, A2A and similar. These are examples, not a fixed list
• Make coding agents, skills and harnesses part of how software is specified, implemented, reviewed and shipped at ZON
• Build internal tools, APIs and harness templates other engineers use every day
• Prototype multi-step and multi-agent workflows (planner/executor, hierarchical, collaborative, swarm, human-in-the-loop) and document when each pattern fits
• Design harness patterns: tool dispatch, session and state, context as a managed resource, memory and compaction
• Explore self-improving loops: traces to eval suites to prompt, skill and tool updates, with regression gates
• Set the standard for how we test AI-built code and agents
• Set up eval approaches the team can reuse: golden datasets, trajectory evals, automated graders where valid, human review where not
• Champion new methodologies and carry them through to adoption: playbooks, demos, pairing and follow-through until the new way is the normal way
• Keep the company current on what’s changing in AI tooling, and say clearly what isn’t worth our time