- Own the vision, roadmap, reliability, documentation, and adoption of the agentic harness.
- Define and run the lifecycle for agents and reusable skills, including building, evaluating, versioning, releasing, observing, and retiring them.
- Build and maintain evaluations for agents and skills to identify performance, regressions, and release safety.
- Write, review, and change code in the harness and its tools, and build prototypes with AI-assisted development.
- Partner with Engineering on APIs, model context protocols, tool access, identity and permissions, data flows, observability, and failure handling.
- Own the internal AI tools portfolio and make build, buy, integrate, automate, or stop decisions.
- Conduct workflow discovery with internal users, map processes and failure modes, and quantify friction costs.
- Work with product leadership on internal skills and agent capabilities that may become customer-facing, including skills delivered through MCP services.
- Partner with UX to ensure the harness produces work that follows Cyderes’ design system and standards.
- Define product and operational metrics, plan rollouts, enable users, and retire or consolidate tools that do not create enough value.
RESTful APIs