Principal Product Manager

Zeta Global•New York, NY

5h•$185,000 - $205,000

About The Position

Zeta Global (NYSE: ZETA) is the AI-Powered Marketing Cloud that leverages advanced artificial intelligence (AI) and trillions of consumer signals to make it easier for marketers to acquire, grow, and retain customers more efficiently. Through the Zeta Marketing Platform (ZMP), our vision is to make sophisticated marketing simple by unifying identity, intelligence, and omnichannel activation into a single platform – powered by one of the industry’s largest proprietary databases and AI. Our enterprise customers across multiple verticals are empowered to personalize experiences with consumers at an individual level across every channel, delivering better results for marketing programs. Zeta was founded in 2007 by David A. Steinberg and John Sculley and is headquartered in New York City with offices around the world. To learn more, go to www.zetaglobal.com.

Requirements

Demonstrated experience leading complex technical products, particularly those involving LLMs, AI agents, workflow systems, developer platforms, ML infrastructure, or AI-driven applications.
Strong understanding of LLM agents, tool use, orchestration, multi-agent workflows, state management, context management, and the architectural patterns required to operate agentic systems reliably at scale.
Experience designing or working with context graphs, knowledge graphs, semantic systems, memory architectures, metadata platforms, or other systems that allow applications and models to understand relationships between users, data, actions, and business objects.
Strong understanding of instrumentation, telemetry, evaluation, and observability for complex software or AI systems. Ability to define the signals required to distinguish between model failures, orchestration failures, context failures, tool failures, and UX failures.
Exceptional ability to align senior stakeholders and cross-functional teams around shared technical architecture, product priorities, ownership boundaries, and operating standards. Comfortable leading initiatives where no single team controls the entire outcome.
Ability to reason across product experience, model behavior, data, infrastructure, APIs, organizational ownership, and operational processes rather than optimizing individual components in isolation.
Strong technical background with hands-on familiarity with tools such as LangSmith and experience working with APIs, workflow orchestration, LLM agent chaining, MCP, evaluation frameworks, and modern AI development environments.
Demonstrated ability to troubleshoot ambiguous technical problems, rapidly prototype potential solutions, and translate experimentation into scalable product and architectural decisions.
Excellent communication skills with the ability to translate highly technical concepts into clear product strategies, operating models, and decisions for technical and non-technical audiences.

Nice To Haves

Experience building or operating agentic infrastructure, AI platforms, context platforms, knowledge graphs, or developer ecosystems.
Experience integrating traditional machine learning with generative models and agentic systems, including using predictive models as tools or contextual inputs for agents.
Experience with workflow orchestration platforms and distributed systems involving multiple services, teams, and execution environments.
Experience designing AI telemetry, evaluation systems, experimentation frameworks, or production observability for LLM-powered products.
Demonstrated success establishing cross-functional technical standards and governance across multiple engineering and product organizations.
Experience building systems where context, telemetry, and evaluation form a continuous learning loop, allowing agent behavior and product experiences to improve based on real-world usage.

Responsibilities

Develop and manage the architecture for chaining LLM agents, tools, models, and workflows across complex use cases. Ensure seamless orchestration, handoffs, state management, and integration across the platform.
Lead the development of a shared Context Graph that gives agents persistent awareness of users, brands, accounts, workflows, capabilities, data, prior actions, goals, and outcomes. Define how context is captured, structured, retrieved, governed, and made available across agents and products.
Implement and manage context streaming services that provide agents with real-time awareness of user actions, application state, system events, and relevant business data. Ensure context remains current, permission-aware, and usable across multi-step workflows.
Define the telemetry framework required to understand how agents operate in production. Instrument and analyze intent routing, agent and tool selection, context utilization, handoffs, latency, errors, completion rates, confidence, user interventions, and business outcomes. Build the feedback loops necessary to continuously improve agent performance.
Build robust evaluation frameworks for testing agent quality, reliability, routing, context utilization, tool execution, and end-to-end workflow completion. Establish both offline and production evaluation methodologies that enable measurable improvements over time.
Lead the creation of a Model Workbench designed for marketers and other non-technical users, enabling them to safely leverage LLMs, traditional ML, agents, and workflows without requiring deep technical expertise.
Oversee the registration, documentation, governance, and discoverability of Model Context Protocol servers, tools, agents, and platform capabilities. Ensure capabilities are easy for both developers and agents to understand, select, and invoke correctly.
Drive alignment across Product, Engineering, Data Science, Design, Analytics, Security, and business stakeholders around shared agentic architecture, context standards, ownership models, evaluation criteria, and platform priorities. Establish clear accountability and operating models for capabilities that span multiple teams.
Define standards for how agents, tools, context sources, telemetry, and workflows are built and integrated across the organization. Balance local team autonomy with the consistency required to create a coherent platform experience.
Actively participate in troubleshooting and debugging using tools such as LangSmith and related observability platforms. Lead by example by rapidly building proof-of-concepts to validate technical approaches, identify architectural constraints, and demonstrate new product opportunities.
Promote a culture of rapid prototyping, experimentation, and evidence-based iteration. Use lightweight development and “vibe coding” where appropriate to quickly turn ideas into working experiences before investing in production-scale implementations.