AI product manager

AI Product Manager Jobs: What the Role Is and Who's Hiring

There are 68 open AI product manager roles across 24 companies tracked by the Agentic Ready Jobs Index, as of 12 June 2026. That is about 8% of the 866 agentic roles in the index, and it is a strikingly senior slice: 67 of the 68 postings are pitched at product-manager level or above, 17 of them at staff or principal. Only one is a mid-level opening. Companies are not hiring product managers to learn AI on the job — they are hiring people who have already shipped a model-backed product and can prove it.

Open roles118
Companies58
Remote share10%

What is an AI product manager?

An AI product manager owns a product whose core behaviour comes from a model rather than from deterministic code. The job carries everything a conventional product role carries — customer discovery, prioritisation, go-to-market — plus a set of problems that only exist when output is probabilistic: defining what "good" means for a system that gives different answers to the same input, deciding which failure rates are acceptable for which users, and trading model quality against cost and latency on every feature.

The practical difference from classic product management shows up in acceptance criteria. A conventional PM writes "the export button produces a CSV"; an AI PM writes an eval — a test set, a scoring method, and a threshold the model must clear before the feature ships. Roadmaps shift too, because the underlying capability changes every time a model provider releases something new, so a feature that was infeasible in one quarter can be table stakes in the next.

The current postings show the role attaching to specific product surfaces rather than to "AI" in general. Datadog is hiring product managers for AI and data security and for database AI optimisation; MongoDB for its "AI Builders Experience"; GitLab for AI custom models; Databricks for AI operations; Stripe for an AI-driven growth outreach motion. The pattern is consistent: the title says AI, but the job is owning one concrete workflow that a model now sits inside.

Skills and tools

The snapshot's postings for this category list no single dominant tool, which is itself informative — the role is defined by judgement rather than by a stack. What the titles ask for, read together: fluency with evals (you will be the person deciding whether the model is good enough to ship), comfort with model cost and latency trade-offs, and increasingly a security dimension — two of Datadog's five openings pair AI with data security. Domain attachment matters more than AI breadth: two of Scale AI's four openings are for public-sector AI applications in Riyadh, Doha, and Dubai, and Databricks wants an outbound PM for AI operations in Singapore. Expect to be tested on a product you have shipped, not on transformer internals.

How to break in

The seniority profile of the postings — 67 of 68 above mid-level — means the realistic route is sideways, not upwards. Product managers add an AI surface to an existing portfolio; engineers who have shipped model-backed features move into agent engineer or AI engineer roles first and pick up product ownership from there. The fastest credibility-builder is a shipped eval: a test set you designed, a threshold you defended, and the decision you made when the model missed it.

Adjacent roles worth comparing: evals engineer (you define quality; they build the machinery that measures it), AI solutions architect (customer-facing design rather than product ownership), forward deployed engineer (shipping inside one customer instead of for many), and agent ops engineer (running what you ship). If your interest leans toward policy and risk, see AI governance lead.

Skills appearing in real postings

EvalsModel cost and latency trade-offsAI and data securityDomain attachmentGo-to-market

Hiring for this role right now

Live from the Agentic AI Jobs Index, updated 16 June 2026.

Salary

None of the 68 tracked postings discloses a salary range. For a public benchmark: levels.fyi reports a median product manager package of $228,250 in the US across all companies and specialisations — not AI-specific. At companies in this dataset, levels.fyi lists Scale AI product managers at $173K–$239K+ and OpenAI product managers at $300K–$950K+, again for all PMs rather than AI-titled roles specifically. Disclosed ranges vary widely by company stage and location.

Sources: levels.fyi — product manager · levels.fyi — Scale AI product manager · levels.fyi — OpenAI product manager

Common questions

Frequently asked questions

How many AI product manager jobs are open right now?

68 open AI product roles across 24 companies, as of 12 June 2026, per the Agentic Ready Jobs Index. The largest single hirer is Sierra with 12 openings, followed by Elastic with 7.

Do AI product manager roles require a technical background?

The postings do not ask for ML research credentials, but 67 of the 68 are pitched above mid-level, and the work assumes you can read eval results, reason about model cost and latency, and argue with engineers about thresholds. Shipped experience with a model-backed product counts for more than coursework.

Are AI product manager jobs remote?

Mostly not. 10 of the 68 tracked postings (15%) are remote — slightly above the 13% remote share across the full index, but the bulk of openings name an office, concentrated in San Francisco and New York.

Hiring for this role — or trying to become the organisation that can use it? See where you stand.

Take the assessment →