Lead Product Manager, Agent Runtime & Evals

Juniper Square•,

7d•$200,000 - $230,000•Remote

About The Position

Juniper Square is hiring a Lead Product Manager to own the agent runtime and evaluation layer of Foundry, its platform for building, deploying, and operating AI agents, apps, and code. Foundry is designed so business specialists can describe an automation and bring it into governed production without an engineering cycle. The role reports to the Group Product Manager for Foundry and partners closely with a dedicated engineering team. Approximately 75% of the role focuses on the agent runtime—including versions, deployments, evaluations, testing, and model releases—and approximately 25% involves partnering with the workflow orchestration PM on the contracts that allow hosted agents to work within governed workflows. The platform is early-stage and involves complex, ambiguous product decisions. The Lead Product Manager will shape the foundations for agentic fund administration, where people and agents build, configure, and deploy automations together. AI fluency is expected, including practical use of modern AI tools to prototype, build and deploy agents, and accelerate product work. The role may benefit from in-person collaboration in San Francisco, but exceptional candidates elsewhere in the United States and Canada are welcome. Juniper Square supports fully remote work as well as work from its offices in San Francisco, New York City, Mumbai, and Bangalore. Success includes enabling real fund administration workflows, such as capital calls and distributions, to run with automations built and deployed on Foundry, followed by rollout to the full fund administration team. The long-term vision is an AI factory in which AI observes work and proposes workflows and agents, while people focus increasingly on oversight and course correction.

Requirements

At least 7 years of product management experience in B2B or Enterprise SaaS, including at least 3 years in an AI-oriented role.
Proven end-to-end ownership of a complex product area, from strategy and roadmap through execution and iteration; zero-to-one experience is preferred.
Hands-on understanding of agent evaluation, including golden datasets, data labeling, automated judges, regression testing, and when human review is needed.
Experience building agents and replacing human-driven processes with products and automation, with evaluation and monitoring discipline in production.
Ability to partner deeply with engineers on topics such as typed contracts between agents and workflows, structured and deterministic outputs, and agent memory.
Product focus on how people use tools, rather than a technical program management focus.
Empathy for non-engineering builders and operators, including users overseeing large numbers of automated agents.
Fluency with LLMs and common AI tools such as Claude and Cursor, with the ability to run prototypes and use AI to accelerate work.
Strong written and verbal communication skills, including the ability to articulate product decisions to engineers, designers, executives, and customers.
Track record of leading through influence and setting strategy that multiple teams execute against.

Nice To Haves

Deep experience with managed agent platforms such as Claude Managed Agents or Workato Genies, including evaluations, tool and connector access, versions, and deployments.
Experience mentoring or managing other product managers.

Responsibilities

Own the vision, strategy, and roadmap for an agent runtime that lets business users take an agent, app, or script from idea through governed production.
Define the automation lifecycle: draft, test, publish, version, deploy, run, observe, improve, and retire.
Define how automations are evaluated before release, including golden datasets, data labeling, regression gates, and decisions about automated versus human evaluation.
Shape evaluation-based model eligibility for agents so model releases can be managed as product decisions.
Partner with engineering on quality checks for agents that evaluate other agents, and make those checks understandable to builders.
Define how agents retain and compact long-lived memory and improve from experience, with attention to adoption and user impact.
Own model routing, cost attribution, and pricing for a model-agnostic platform.
Define how hosted agents access data and take actions across Juniper Square products, systems of record, and third-party tools.
Own reusable connector, credential, and access-grant models, as well as agent identity and run modes.
Design fleet oversight capabilities, including dashboards, run history, observability through traces, metrics and logs, kill switches, and auditability.
Partner with the workflow orchestration PM on the agent registry, typed contracts, and workflow bindings; ensure agent-driven steps stage changes for human release.
Partner with design and engineering to build tools that let business SMEs configure, test, and tune automations, favoring convention over configuration.
Track managed agent runtimes, agent frameworks, and automation platforms, and make clear build, buy, and partner recommendations.
Set direction across the platform and adjacent teams, raise product craft standards, and guide engineering through ambiguous platform tradeoffs.