Analysis of 651 PM roles at 36 AI companies: AI PMs earn $262K median, most don't need ML experience

aakashgupta · x · 2026-10-10

The author ran an AI model over 8,354 open roles at 36 AI and tech companies (OpenAI, Anthropic, xAI, Databricks, Perplexity, Cursor, Figma, Stripe and more), filtering down to 651 unique product/PM roles.

Key findings:

151 AI PM roles don't require hands-on ML experience; only 15 do
Where pay is listed, no-ML AI PM roles have a median of $262K vs $216K for non-AI PM roles
Only 8 of 651 roles are entry level (APMs, new grads, interns)
Anthropic's Claude Tag PM role paid $385K–$460K and explicitly said "an engineering background is not required," but wanted a strong grasp of model capabilities and fluency working with engineering teams
OpenAI's Core Models PM posting asks for pairing offline evaluation with online experimentation and user signals

Takeaway: the real bar is knowing what models can do and how to measure it — both learnable without an ML degree. The scan used Jev, a model that answers multiple-choice questions about data, reading all 781 postings in 6.4 seconds for 7.6 cents, and correctly caught cases like Datadog's "APM" (Application Performance Monitoring) that keyword filters get wrong.

Original post →

Analysis of 651 PM roles at 36 AI companies: AI PMs earn $262K median, most don't need ML experience

More from Companies & People