Description

Own the architecture, development, and production operation of multi-agent AI systems and data platforms.

This role is on-site.

Responsibilities

• Design and build agentic AI systems that serve live business workloads, owning supervisory orchestration, tool calling, state management, and error recovery for long-running workflows.
• Develop agent registry platforms with sandboxed execution, automated validation of generated artifacts (including ingestion code and DBT packages), and engineer review-and-approve deployment gates.
• Operate LLM observability in production, implementing per-run tracing of agent decisions, token and cost monitoring, and triage of agent failures.
• Build and maintain agent evaluation harnesses, including offline eval suites and regression testing that enforce quality gates to block releases on failure.
• Implement guardrails and human-in-the-loop controls, including approval gates, permission-scoped tool access, audit logging, and rollback procedures.

Required Skills

• 12+ years in software or data engineering, including 3+ years building LLM-based systems and 2+ years designing/operating agentic AI systems in production.
• Proven track record as lead architect of at least one multi-agent system running in production for 12+ months.
• Strong proficiency in Python and SQL.
• Experience building or deploying MCP servers/clients or OpenAI-compatible tool interfaces in production systems.
• Hands-on experience with CI/CD for data platforms and ETL/ELT processes.
• Deep knowledge of data platforms, including DBT, SSO (SAML/OIDC), IAM, and RBAC design.

Key Skills

Education

Bachelor's degree