AI Developer Advocate Job Description

The Developer Relations Foundation defines developer relations as "the practice of building and nurturing relationships with external and internal teams through community engagement, technical support, education, and advocacy to enable the successful adoption of an organization's developer products and drive business value." An AI Developer Advocate does that work for a model provider, an AI API, an agent framework, a vector database or a cloud AI service. The customer is a developer with an editor open, and the job is to get that developer from an API key to a working feature with as little friction as possible.

The outward half: code and teaching

The visible output starts with code. An advocate builds sample apps that call a chat or embeddings endpoint, wires up tool calling and structured output, shows streaming responses in a web UI, and keeps quickstarts in Python and TypeScript passing against the current SDK release. They publish Jupyter notebooks for retrieval pipelines, write reference MCP servers that expose a database or a SaaS app to an agent, and maintain starter repositories on GitHub that people fork as the first commit of a real project.

Around the code sit the teaching formats: tutorials, blog posts, short videos, livestreams, workshops and conference talks. GitLab's Developer Advocate job description, a published example of the role at a developer-tools company, says advocates "might write sample code, author client libraries, provide insights to journalists," and lists blog posts, demos, videos and technical presentations among the content they create. On an AI platform the same list applies, with the subject matter shifted to prompts, context windows, rate limits, evaluation and agent design.

The inward half: feedback to product

The other half of the job happens inside the company. An advocate spends hours in Discord servers, forums and GitHub issues, reproduces the bug a developer reported, and turns a pile of complaints into a ranked list the product team can act on: the SDK method nobody can find, the error message that hides a quota problem, the tutorial that breaks when a model version is retired. Before a launch they act as the first outside-minded tester, running the new endpoint the way a stranger would and flagging the gaps in the getting-started path.

Who they work with

GitLab's description names marketing, engineering, product, business development, sales and partners as the advocate's daily contacts. In an AI company those contacts include SDK and API engineers, documentation writers, product managers for the model or platform, the events team, and partner engineers at companies building integrations.

What it is not

An AI Developer Advocate is not a sales engineer. A sales engineer prepares demos and proofs of concept for one prospect's purchase and works toward a quota. An advocate works in public for every developer at once, and the measure is adoption and developer success rather than a closed deal. The role also differs from a technical writer: advocates write, but they also build, speak, run community programs and argue for product changes.

No license or degree is legally required. As a rough guide, people move into AI developer advocacy after about two to five years of building software, for example from backend, full-stack or machine learning engineering, and most start out with a degree in software engineering, computer science or a nearby field, usually at the bachelor's level. A public portfolio can carry a candidate without the degree: repositories other people use, merged pull requests to open source AI libraries, recorded talks and tutorials that rank in search.

Technical skills

The technical bar is a working developer's bar. Expect to write production-quality Python and TypeScript, call REST and streaming APIs, handle authentication and retries, and debug other people's code quickly. AI platform work adds its own vocabulary: tool and function calling, structured outputs, embeddings and retrieval, context management, prompt design, evaluation harnesses, and agent frameworks. Familiarity with the Model Context Protocol and with AGENTS.md conventions for coding agents is useful because advocates build and document integrations that rely on both.

Communication skills

The other bar is teaching. Advocates explain a concept in a five-minute lightning talk, then again in a long written guide, then again in a forum reply to someone who is stuck. GitLab's published job family is a concrete example of what one employer asks for: "Experience creating software or contributing to open source software" and "at least 1 year of experience giving talks and developing demos, workshops, webinars, videos, and other technical content." The same description asks for the ability to manage a conference schedule with its CFP deadlines and show dates.

Credentials

Certifications are optional and secondary to shipped work, but a few map directly to the job:

• Model Context Protocol Associate (MCPA), from the Agentic AI Foundation, formed under the Linux Foundation. It was announced Sept. 14, 2026 as a 120-minute, online, proctored, multiple-choice exam covering five domains, including security and governance.
• AWS Certified Machine Learning Engineer - Associate (an Amazon Web Services credential), for advocates whose platform runs on or integrates with AWS.
• Google Cloud Professional Machine Learning Engineer, from Google Cloud, for the equivalent on Google's platform.

Getting in

Entry routes include moving from an engineering team at the same company, from a solutions or support engineering role, or from an independent track record of open source work, meetup organizing and technical writing. A practical first step is to pick an AI API, build three small but complete projects with it, write them up clearly, and give a talk about one of them at a local meetup.

BLS does not publish employment or projections for developer advocates, so the nearest official figures come from the broader software group. The Occupational Outlook Handbook projects employment of software developers, quality assurance analysts and testers to grow 10 percent from 2025 to 2035, "much faster than the average for all occupations," with about 106,100 openings a year on average across that group. Those numbers describe engineers broadly, not developer relations teams, and they should be read as background rather than as a forecast for this title.

The standards layer moved to neutral foundations

One change for AI developer advocates since late 2025 is what they are advocating for. On Dec. 9, 2025, the Linux Foundation announced the formation of the Agentic AI Foundation (AAIF), with founding contributions from Anthropic, Block and OpenAI: the Model Context Protocol, the goose agent framework, and AGENTS.md, which the foundation describes as a standard that gives AI coding agents "a consistent source of project-specific guidance." For an advocate at a platform company, part of the job can include teaching open, vendor-neutral protocols alongside the company's own SDK, and contributing sample servers and examples to shared repositories as well as to the company's own.

Documentation is still the front door

The 2026 Stack Overflow Developer Survey, a global poll that drew 30,903 responses from 169 countries, found that technical documentation (59%) was the leading way respondents learned to code, edging out AI codegen tools (53%). The same survey found that around 79% of developers were focusing on AI skills in their learning. For an AI Developer Advocate, those two findings point at the same deliverable: clear, accurate docs and examples for AI features, written so that a person reading them and an assistant quoting them both get the call right.

Career paths

Developer advocacy has its own ladder at companies that publish one. GitLab's job family lists Developer Advocate, Senior Developer Advocate, Staff Developer Advocate and Principal Developer Advocate levels, alongside Developer Relations Program Manager and Director of Developer Relations roles. Advocates also move sideways into product management, where their record of developer feedback is direct evidence of user insight; into developer experience or SDK engineering; into solutions architecture; or into partner engineering for an AI platform's integration ecosystem.

What to watch

Where the team sits matters. At GitLab, for example, the Developer Advocate job family is filed under marketing roles, so it is worth asking in interviews where the team reports, how its work is measured, and whether the role owns any product or documentation surface outright. Keep a running record of shipped code that other developers depend on and of product changes that came from your feedback; it makes the case for the role concrete. The work also uses the same AI tools it teaches: an assistant can draft a tutorial or a sample app, so the parts to build a career on are the ones a draft cannot supply, such as testing what actually runs, catching inaccurate claims, and keeping the trust of the developers you serve.

Dear Hiring Manager,

I am applying for the AI Developer Advocate position on your developer relations team. For the past several years I have worked as a backend engineer on a payments product, and for the last two of them I have spent my evenings and weekends doing the work this role describes: building with AI APIs in public, writing about what broke, and helping other developers get unstuck.

The project I am proudest of is an open source starter kit for retrieval-augmented support bots. It began as a weekend experiment with your embeddings endpoint and grew into a repository that other developers fork as the first commit of their own projects. Maintaining it taught me what developers actually trip over: authentication errors that look like model errors, SDK upgrades that silently change defaults, and quickstarts that assume a reader already knows what a context window is. I rewrote the README three times based on issues people filed, and each rewrite cut the number of repeat questions.

I also speak. I have given talks at two regional Python meetups and a community conference on building tool-calling agents, and I run a monthly online workshop where attendees build a small agent from scratch in ninety minutes. Recording those sessions and turning the questions into written guides is my favorite part of the work.

On my current team I am the person who files the most detailed bug reports against our vendors, with reproduction steps, logs and a suggested fix. I would like to bring that habit inside an AI platform, where turning developer frustration into a clear product request is part of the job rather than a side effect.

I would welcome the chance to walk you through the starter kit and talk about how I would approach your next SDK launch.

Sincerely, Jordan Avery