Principal Product Manager, Agentic AI

Role

Principal Product Manager leading the design, prototyping, evaluation, and enterprise deployment of agentic AI for workforce management. This is an individual-contributor role that requires hands-on product work across specification, implementation review, evaluation, and commercial readiness.

Key Responsibilities

• Define agent responsibilities, tool integrations, escalation boundaries, and manager-facing surfaces with attention to pay and legal compliance.
• Select and prioritize workflows suited for agents through operational research with customers (district managers, store managers, frontline employees).
• Specify human-in-the-loop contracts: autonomy thresholds, escalation rules, override behavior, logging, and auditability.
• Stay engaged through implementation: review builds against specifications, join bug triage and design reviews, and hold releases when quality is insufficient.
• Build and maintain evaluation sets and pass criteria; review failure traces and hold releases when evaluation is not persuasive.
• Produce and maintain Functional Requirements Documents (FRDs) that serve as ground-truth specifications.
• Partner with go-to-market on packaging, metering, and pricing inputs for agentic products.
• Lead design-partner pilots for large enterprises: scope pilots, set success criteria, and determine GA vs bespoke work.

Technical Expectations

• Write and iterate on prompts and system prompts; explain behavioral improvements from changes.
• Stand up working prototypes of agents (tool definitions, structured outputs, retries, guardrails) before engineering builds.
• Use agentic coding tools (e.g., Claude Code) daily for prototyping, data analysis, and specification work.
• Diagnose execution traces to determine root causes (tool description, retrieval gaps, context-window constraints, output schema, or model limitation).
• Build and run evaluation sets and interpret results without a dedicated data scientist.
• Hold informed positions on retrieval design, tool-use patterns, structured output, model routing, latency/cost trade-offs, and multi-tenant safety in enterprise SaaS.
• Open pull requests for prototypes, prompts, evaluation fixtures, or specification stubs when appropriate.

Requirements

• 7+ years in product management (Senior) or 10+ years (Principal), including at least 2 years shipping LLM-backed features to external, paying users.
• Demonstrated hands-on capability (expect to show work and prototypes).
• Substantial enterprise B2B SaaS experience (multi-tenant architecture, RBAC, audit trails, SSO, data residency, security review, procurement).
• Track record taking ambiguous, operationally complex domain problems through to shipped products with demonstrated adoption.
• Evidence of owning product quality through implementation and release decisions.
• Judgment about current model limitations and willingness to state limitations clearly.
• Comfort reading and working inside a codebase; ability to run code locally and open pull requests.
• Strong written communication and documentation skills.

Preferred Qualifications

• Experience in workforce management, HCM, ERP, supply chain, or other operationally complex, compliance-heavy domains.
• Familiarity with labor compliance regimes or ability to become fluent quickly.
• Background in demand forecasting, optimization, or decision systems.
• Experience building configurable platforms or authoring customer-configurable surfaces.
• Engineering or technical degree or prior experience as an engineer or data scientist.
• Experience with products sold into frontline-heavy industries (retail, hospitality, healthcare, manufacturing).

Scope

• Individual contributor role with no direct management responsibility.
• Accountable for what ships; specification is a part of the work but not the whole.
• Focused on product definition, prototyping, specification, evaluation, and delivery rather than AI vendor strategy or vendor selection.
• Requires independent judgment developed from operational research in ambiguous, rapidly changing model/agent landscape.

Selection Process

• Includes a practical exercise: design an agent for a sanitized workforce management problem and deliver a working prototype; Legion compensates candidates for the exercise time.

Compensation & Benefits

• Salary range: $280,000 - $300,000 plus bonus (tied to business outcomes) and stock equity.
• Benefits include fully remote work, medical/dental/vision with $0 monthly premium options effective first day, 401k, discretionary paid time off and paid holidays, parental leave, equity, monthly wellness reimbursement, and a monthly lunch stipend.