Agent Experience (AX) Specialist at Cloudinary – Kraków
Explore Related Opportunities
About This Position
AI agents (Claude Code, Cursor, Windsurf, GitHub Copilot) are becoming a primary way developers discover and integrate tools. When a developer asks an agent to add image and video management to their app, the agent’s autonomous decision to install a tool, connect to its API, and start using it is the new top-of-funnel.
We call this motion Agent Experience (AX): distinct from SEO (human finds us via search engine) and AEO (human is influenced by an LLM recommendation). AX is when the agent itself discovers, evaluates, and uses a tool without the human initiating each step.
You will be the first marketing person dedicated to this motion at Cloudinary. Your job is to make AI agents independently discover, recommend, and use Cloudinary, and to build the measurement infrastructure to prove it’s happening.
This is not a strategy role. You ship weekly across four workstreams:
MCP Infrastructure: Optimize how agents understand and interact with Cloudinary’s MCP tools. You’ll own tool metadata, error responses, registry presence, and package discoverability.
Agent Testing & Benchmarking: Systematically test how major agent platforms recommend Cloudinary vs. competitors. Track inference share over time and identify gaps.
Dashboard & Measurement: Build and maintain the reporting infrastructure that ties agent activity to registrations and product usage. Define the AX funnel and own the numbers.
Agent-Facing Content: Create and maintain pages, structured data, and example repos optimized for agent consumption—not human browsing.
These are non-negotiable. You cannot do this job without them.
1. MCP & Agent Ecosystem Fluency
You understand what MCP (Model Context Protocol) is, how MCP servers work, and how agents discover and connect to them.
You can read and edit MCP tool descriptions, JSON schema configs, and registry metadata without engineering support.
You have hands-on experience with at least 2 major AI coding agents and use them in your own workflow.
You understand how agents decide which tools to call—tool descriptions, parameter schemas, and error responses are your optimization surface.
2. Technical Comfort (APIs, CLIs, Logs)
You can write and run queries against structured server logs to pull connection and usage metrics.
You are comfortable with npm: publishing packages, updating metadata, reading download stats.
You can use Git and GitHub: clone repos, edit files, open PRs, create example repos.
You can work with REST APIs, understand OAuth flows, and read/write JSON configurations fluently.
3. Analytics & Measurement
You can set up and maintain analytics custom reports, configure events, and build funnel analyses.
You understand attribution modeling, especially the challenge of attributing signups to agent interactions that don’t generate traditional web traffic.
You can build and maintain a tracking dashboard. Data discipline matters more than tooling.
You think in funnels: discovery → connection → usage → registration → adoption.
4. Growth Mindset & Experimentation
You naturally think in A/B tests: which copy, which metadata, which structure drives better outcomes?
You can design, run, and interpret experiments without a data science team.
You ship weekly. Strategy without execution is a document no one reads.
5. Content for Machines, Not Humans
You understand how agents consume content: structured data, direct answers, code examples up front.
You can write technical documentation optimized for agent parsing—clear headings, explicit capability statements, comparison callouts.
You know the difference between content that ranks in Google (SEO) and content that gets cited by agents (GEO)—and can create both.
Experience with media management APIs (image/video upload, transformation, delivery).
Familiarity with GEO tools: Profound, Otterly.AI, Ahrefs Brand Radar, or similar.
Prior experience in developer relations, developer marketing, or developer tools growth.
Understanding of the competitive landscape in cloud media management.
Experience building or contributing to MCP servers.
Working knowledge of Python scripting for data analysis and automation.