Principal Product Manager, Agentic Observability and Governance

Description

The company uses automation to change how work gets done and builds category-leading enterprise software to put that power in customers' hands.

The team wants people who are curious, self-driven, generous, and genuine, and who enjoy a fast-moving, growing company where people look out for each other and the wider purpose.

This product manager will help advance the Agentic Automation vision so enterprises can observe, trust, and govern every agent they run, whether it runs on the company's platform or elsewhere.

The scope covers three connected areas. Traces and observability record what an agent did, what it cost, which model and configuration it used, and why. They serve as the single source of truth across the platform, including agents and every service connected to the trace infrastructure. Agent identity defines who owns an agent, whom it acts for, and what it may do. Governance links the two and connects the Agents team to platform-wide governance.

You will help define patterns the industry is still working out: OpenTelemetry for agents, cross-platform trace exchange, delegated agent permissions, and policy and audit for non-deterministic systems.

The role is hands-on and AI-native, with no status meetings. You read traces, prototype with coding agents, speak with customers and FDEs, and turn findings into clear direction for engineers. Your main lever is influence rather than team size.

The goal is that every production agent is visible, attributable, and governable. Builders can debug a failed run in minutes, finance owners can trust cost figures, and security teams can establish what an agent did and for whom without filing a ticket.

Responsibilities

• Own the schema, completeness, and quality of every trace, including the accuracy of cost, token usage, and LLM configuration data.
• Help shape the multi-year vision for agentic observability and governance across the company's platform and the external agents it works with.
• Make sure every agent and every run maps to an owner, the user it acts for, and the permissions it used.
• Connect traces and identity to policy, audit, alerting, and controls such as a kill switch, acting as the Agents team's counterpart to platform governance.
• Build integrations with other platforms, including OpenTelemetry export and import, Microsoft Agent 365, and other agent and observability ecosystems. Company traces flow out and outside agent traces flow in.
• Bring company services and teams onto the trace infrastructure so traces cover the whole platform, not only agents. Lead intake, prioritization, and coordination of trace requests across the company.
• Work with the largest and most demanding customers, including those in regulated industries, on production deployments. Dig deep enough to identify which missing span mattered and why.
• Build with the platform yourself: deploy agents, read their traces, connect exports to real observability tools, and prototype with coding agents. Where a point needs proof, build a working example.
• Turn the fast-changing OpenTelemetry GenAI conventions, agent observability vendors, agent identity standards, and AI regulation into one coherent company point of view and a shipping product.
• Set the metrics that count, such as trace coverage and completeness, cost accuracy, time to diagnose a failed run, and the share of agents with a clear owner and identity, then prioritize hard against them.
• Present the company's agent observability and governance story to customers, partners, and industry forums.

Requirements

• 6+ years in technical product management or engineering, focused on observability, developer platforms, identity, security, or applied agentic AI.
• A track record of shipping infrastructure that builders and platform teams depend on.
• Hands-on experience building observability, telemetry, tracing, or governance products that capture, correlate, and act on signals across many services.
• Strong command of agentic AI: models, tool calling, MCP and A2A, coding agents, agent design patterns, evals, and observability. You can read a trace, spot what went wrong, and see how the trace could be improved.
• Working knowledge of identity and access concepts such as service identities, delegated permissions, OAuth, and audit, with a view on how they apply to agents.
• A builder's mindset: you write code or direct agents to write it, test ideas, and ship prototypes instead of writing specs and waiting.
• First-principles thinking and comfort with ambiguity. You reach a defensible direction in days, not quarters.
• Clear communication: you write position docs that prompt bold decisions, explain technical detail to skeptical architects, and run demos that win over a room.
• Comfort working in regulated, on-premises, and hybrid environments.
• Ideal background: has built observability, telemetry, identity, security, or agent platform products at a frontier lab, hyperscaler, observability vendor, or automation vendor, or was an early PM at a startup in the space.
• Ideal background: has taken a product from 0 to 1. Startup or founding PM experience is strongly preferred.
• Ideal background: holds a well-evidenced view on how enterprises will observe and govern agents over the next 24 months.

Conditions

The work model varies by team. Flexibility applies to when and where work gets done, and the mix of hybrid, office-based, and remote workers depends on business needs and the role.

Applications are reviewed on a rolling basis with no fixed deadline. The window may change with application volume, or may close once a qualified candidate is selected.