AI Engineer

Uppgiften

Embed with the teams doing the work, map how a process actually runs, and identify the single highest-leverage opportunity to automate. Build tool-calling agents on our platform that take real processes off people’s plates as reliable production systems, not prototypes. Turn every build into reusable skills, connectors, evals, and templates, so agent-building gets faster each time. Hand off cleanly, leaving a running system with a runbook, an eval framework, and a named owner. Define what “good” looks like for agents — accuracy, cost, guardrails, sensitive-data handling — and mentor the per-team champions who scale it.

Vad vi förväntar oss

• A production software engineer first, with strong Python and real fundamentals in testing, APIs, version control, and observability.
• Proven experience building agentic workflows, RAG, and tool-calling systems that run in production, not just in demos.
• Evaluation-driven: you set a baseline, define the metrics that matter, and prove with numbers that a change moved the needle.
• Fluent in GCP and Vertex AI, comfortable in the Claude ecosystem (Claude Code, MCP, skills), and at home in BigQuery.
• Forward-deployed by temperament, thriving in ambiguity and able to translate business problems into technical plans for a director and an engineer alike.
• Approximately 5+ years in software engineering with production AI experience, strong product sense, and the judgment to know when to prototype, harden, or stop.