We are looking for an experienced software engineer who understands AI & ML pipelines and has shipped models into production.
Everything we build depends on understanding how AI agents read, cite and recommend: which brands get cited, how well an ad fits a page and the question an agent is answering, and what an impression to an agent is actually worth. You’ll own the models behind all of it, from our citation engine and brand extraction today to the models that will match and price ads in our auction tomorrow.
You’ll work closely with our CTO, and own our AI & ML roadmap, the data pipelines your models learn from, and the infrastructure they run on. Knowing what good looks like, and proving it with measurement, is what we’re hiring for.
Our structure is intentionally flat, with influence earned through skill & context. Our codebase is shared rather than siloed, and we move fluidly across priorities as a team. We have dozens of agents working alongside us every day, so we go from an idea to a measured experiment in hours, and from a proven result to production just as quickly. It’s how a small team builds at the pace of one ten times our size.
Who you are:
- You’ve shipped models into live systems where speed and reliability matter, and you know the difference between a good demo and a good product.
- You measure before you believe. You build the evaluation first and trust numbers over impressions.
- You’re commercially minded, prioritising what moves revenue rather than what’s interesting to build. You’ll hear directly from brands and publishers on what your models are worth to them.
- You’re comfortable in a small, high-trust team, happy to pick up whatever’s needed.
- Python & TypeScript fluency preferred.
- Experience in ad-tech, search ranking or recommendations is a strong plus.
Must haves:
- 5+ years of shipping production software, including ML models and the pipelines behind them.
- You work agentic-first. Agents speed up your loop from idea to experiment to production; your judgement decides what’s worth testing and what goes live.
#J-18808-Ljbffr
