AI Content Engineer

Data quality is the bottleneck of the AI era. Every Fortune 500 is trying to solve it. We're the team they call.

Soda is the data quality layer for Disney, Ralph Lauren, CBRE, HelloFresh, 2K Games, and Nubank. Our open-source engine, Soda Core, is used by companies like Tesla, Slack, Adyen, Walmart, and JPMC. The category is growing fast because AI doesn't work on bad data.

We're hiring one person to own content at Soda, end to end.

Soda's mission is to monitor the world's decisions. We're building the first Data & AI performance monitoring platform that spans data collection to automated decision-making. Soda helps customers find, understand, and fix data quality issues.
The Role

The website. Technical blogs, webinars, and product videos. Your own content and following.

One person cannot hand-make all of that. So you'll build the system that can. That isn't a stretch goal or a second phase - we don't think the job exists without it.

You'll join the Growth team reporting to a co-founder.

What you might call yourself.  Content engineer, technical content marketer, growth engineer, marketing engineer, GTM engineer, developer advocate, technical writer who got bored, or nothing in particular. We've seen this job posted under every one of those. We don't care what's on your CV. You need to be able to write and you need to be able to build, and almost nobody can do both.

Why this is a real problem and not a content calendar.  The expertise is already in the building. Founders who have argued about this category for a decade. Customer engineers who spend their week inside broken pipelines at Fortune 500s. Customers who will tell you things nobody has published anywhere. Almost none of it gets written down, and what does get out is hand-made, one piece at a time, by people who have other jobs. Your job is to get it out of their heads and onto the internet, at a volume that is only possible with a machine behind it.

Taste is the bar.  Anti-slop is the whole point. An engine that ships slop faster has negative value, and most of them do. We're hiring for the judgment to look at an accurate, competent, lifeless draft and bin it - then work out what the pipeline did wrong and fix it. The engine has to write in distinct voices: a founder sounds nothing like a customer engineer, and the output should sound like them, not like a model. Your own profile and following is part of how we'll know it works. If it can't make  you  sound good, it won't make anyone else sound good either.

What this is not.  You'll write code every day, but this is not a software engineering job. We're not hiring you to build product, and if you're a backend, data, or platform engineer looking for your next engineering role, this isn't it. It's also not a traditional marketing or DevRel role: no agency to brief and no conference circuit.

Requirements
- You can write, and you can tell good writing from bad : send us something you wrote that a technical audience actually read. The most important line in this ad.
- Serious prompt engineering : multi-step LLM pipelines with structured outputs and evals, not chatting with a model. You know why the naive version produces slop.
- 3+ years building things that go to market : technical content, growth engineering, marketing engineering, GTM engineering, DevRel, marketing ops, RevOps, or a founder background. Nobody agrees what to call this job yet. Titles vary; portfolios don't.
- Enough code to be autonomous : Python or TypeScript, APIs, webhooks, a database. Nothing should be blocked waiting for an engineer. We build agentically here - Claude Code, MCP servers, and agent skills are the daily workflow.
- Enough data fluency to be credible : you're writing for Staff Data Engineers and CDOs. You can hold your own on pipelines, data quality, and warehouses.
- Generalist instinct : you find the constraint and fix it, and you'd rather build the workflow than do the task fifteen more times.
- Independence : you