Product Manager, Enterprise Core Platform

Scale AI•New York, NY

1d•$205,600 - $300,000•Remote

About The Position

Scale is building the agentic operating system that powers AI deployment across the world's most important organizations. The Platform PM is the person who determines what that foundation needs to be — and holds the bar on how well it gets built. This is not a roadmap PM, a program manager, or a feature tracker. The Platform PM owns the core infrastructure that every FD team builds on top of: defining the foundation before anything else can be built reliably, sequencing work against real customer commitments and FD team velocity, identifying which FD-built patterns are ready to graduate to core, and ensuring that every capability that ships does so at a bar customers can trust without thinking about it. The best Platform PMs have defined platform capabilities at production bar in high-stakes environments and understand the difference between what is necessary to build, what is interesting to build, and what should never be built at all.

Requirements

6+ years in product management with meaningful time owning platform, infrastructure, or developer-facing products at production bar.
Demonstrated record of defining and shipping platform capabilities — not features on top of a platform someone else built. You can describe the architectural decisions you made, what production required, and what you learned.
Strong sequencing judgment: can prioritize across competing customer commitments and team velocity needs with incomplete information, without thrashing.
Technical fluency sufficient to hold real conversations with platform engineers about architectural tradeoffs across infra, auth, observability, deployment, and agent runtime. Not deep coding, but genuine comprehension of how production systems work.
Extreme ownership and follow-through: closes loops without reminders, drives outcomes across teams without formal authority, holds self and others to commitments.
Clear, precise communication: adjusts for audience (platform engineer vs. FD PM vs. executive); writes and speaks at the right level of detail.
Platform foundation ownership: has defined what a platform capability needs to be before it can be used reliably in production; understands what 'done' means at the platform layer vs. the application layer.
Sequencing and prioritization judgment: sequences work against real customer commitments and FD team velocity needs — not what is technically interesting or easiest to build.
Pattern recognition: identifies when field-built work has genuinely repeated and is ready to graduate to core; does not rush graduation or delay it out of caution.
Quality discipline: every capability earns its place twice — by being necessary and by being done well enough that it is trusted without thinking about it.
Cross-functional influence: drives field teams and platform engineering toward shared outcomes through evidence and framing, not process.
Operating in ambiguity: creates structure where there is none; makes sequencing calls with incomplete information; does not wait for fully-formed requirements.

Nice To Haves

Direct experience with AI/ML platform infrastructure — agent frameworks, eval pipelines, fine-tuning workflows, or observability for production AI systems.
Experience deploying into constrained environments: government, regulated industries, air-gapped or classified infrastructure.
Prior experience as a platform or infrastructure PM at a company where field-built patterns were a key signal source for the core product.

Responsibilities

Define the foundation. Determine what needs to be true at the core platform layer — across each capability pillar — before anything reliable can be built on top of it.
Sequence it. Prioritize and order platform work given customer commitments, FD team velocity needs, and what is actually blocking progress today. Getting this wrong means building on an unstable base or building the right things in the wrong order and paying for it in rework and FD team drag.
Identify patterns and graduate capabilities. Watch what FD teams are building across the application layer, determine what is genuinely repeating, and make the call on what moves to core — and when. Do not rush graduation or delay it out of caution.
Hold the quality bar. Every capability on the platform roadmap earns its place twice: once by being necessary, and again by being done well enough that customers trust it without thinking about it. Keep the platform from becoming a graveyard of half-built capabilities.
Unlock delivery team time to value. Forward deployed delivery teams should spend their time solving the hardest customer problems, not rebuilding plumbing on every engagement. That requires a platform that is secure by default, deployable anywhere, observable in production, and equipped with the agentic primitives, eval infrastructure, expert judgment capture, and data reasoning capabilities that every agent workload needs.
Partner across the organization. Work closely with platform engineering, forward deployed PMs, and BU leads to surface signal, resolve sequencing tradeoffs, and keep the roadmap grounded in what is actually blocking work.