Senior AI Product Manager, Cortex
An AI-native product builder who turns uncertain ideas into testable products and commercial evidence.
About Cortex
The entry point for consumer purchase decisions is shifting from the search engine to AI answers. Brands are discovering they’re not in ChatGPT or Perplexity’s answers — without knowing why, and without any controllable channel to fix it. Publishers, meanwhile, watch their readers and ad revenue get intercepted by AI.
Cortex is a “brand-in-AI-answers” system. Through an AI Q&A widget embedded in publisher sites, when a reader asks an AI a question inside a content context, Cortex surfaces the right brand in the answer in a way that’s useful and trustworthy — publishers keep the first-party interaction and earn a revenue share, and brands gain a distribution channel into AI answers.
Most GEO tools stop at the measurement layer (diagnosing a brand’s visibility in AI answers). What Cortex builds is the distribution layer: which Q&A contexts a brand should appear in, how it appears credibly, and whether that appearance actually creates value.
The product is already live in market, validating with real brands and publishers. The architecture and business model are still evolving fast — you’re not inheriting a settled system, you’re validating alongside the team what kind of product experience and business model can form a repeatable, scalable market.
Your mission
Own Cortex end to end — from problem exploration, product definition, and prototype validation through to delivery, adoption, and commercialization. Turn ambiguous market signals into testable hypotheses, actionable prototypes, clear priorities, and a roadmap engineering can execute.
This is a highly hands-on role. Faced with a new product hypothesis, you can use AI coding tools, LLM APIs, automation, and no-code/low-code tools to build a working prototype yourself and put it in real users’ hands for evidence — without waiting on the engineering queue, and without needing it to be production-grade. You own the first mile from idea to testable experience.
What you’ll do
- Build Cortex’s product narrative — define the ICP, pain points, differentiation, and business model on both the brand and publisher sides, so engineering, sales, and customers align around the same story.
- Drive continuous discovery — interview customers, form hypotheses, design minimal validations, and find the intersection where brands will pay, publishers will deploy, and readers will engage.
- AI-native prototyping — use tools like Cursor and Claude Code to wire LLM APIs to data, quickly build working prototypes, and get real feedback.
- Break down ambiguous goals — turn a target like “raise POC conversion rate” into executable, testable questions and milestones, and keep moving even when information is incomplete.
- Own prioritization — sequence across brand needs, sales opportunities, competitive pressure, and technical debt; make the trade-offs explicit; and revise openly when new evidence arrives.
- Define AI product specs — understand the trade-offs of RAG, agents, and model routing; define success conditions, failure modes, fallbacks, and evaluation; and talk to engineering directly, not through a layer.
- Build the measurement loop — from reader interaction to publisher adoption to brand outcomes to revenue signals, so “is this product worth the investment” can be answered honestly.
Who we’re looking for
- 5+ years in product management or an adjacent role, having independently owned a product line from discovery through outcome validation.
- At least one stretch of real work on an AI/LLM product — you can read technical architecture docs and discuss trade-offs directly with engineering.
- You can independently build a working prototype with AI coding tools, LLM APIs, or no-code tools, and have used one in customer interviews, usability tests, or POCs.
- You’ve broken ambiguous goals into testable hypotheses and pushed, adjusted, or stopped based on evidence.
- Basic product data analysis and the ability to pull your own data (self-service).
Nice to have: you can build and debug in JS/TS/Python; you’ve implemented RAG, a tool-calling agent, or an evaluation pipeline; you’ve used tools like Cursor, Claude Code, or n8n; you have AdTech / MarTech / media / AI platform or B2B SaaS experience; you’ve been part of a commercialization journey from POC to paid renewal.
You might not be a fit if you
- Mostly relay messages between the demand side and engineering, and rely on others to form the product judgment.
- Can only express product through slides or wireframes, and believe all prototypes should be built by engineering.
- Are great at building AI demos but can’t articulate which product hypothesis they validated.
- Stall without a complete set of requirements, and treat ambiguity as someone else’s failure.
Why now
- A market taking shape. GEO / AI answer visibility is a structural shift happening in the marketing industry — and you’ll stand at the distribution layer, closer to the value exchange than the measurement layer.
- Enormous product leverage. A small, senior team means your judgment directly sets product direction — no layers of approval.
- AI-native ways of working. We use AI tools as part of daily work — your validation speed is the team’s learning speed.
How to apply
Send us something that shows how you think — a product you shipped, a prototype you built to test a hypothesis, a discovery story where the evidence changed your mind. We care about what you’ve shipped more than where you’ve worked.
📧 [email protected]