Senior Product Engineer - Agentic Workflows

Hi, we're Prismic 👋 and we're designing the future of the web.

We believe in autonomous websites and deeply personalized digital journeys. Our mission is to empower marketers and developers to create empathic web experiences at scale and with soul.

We're pioneering an entirely new category: the Automation Platform for Websites. It's what comes after the CMS. Imagine a site that knows your brand by heart and grows itself, continuously optimizing layout and interactions—while staying beautifully on-brand.

🚀 Used by Builders at Bershka, AXA, Deliveroo, TicketMaster

Over 5,000 companies trust Prismic, including 300+ enterprise teams who build and scale with us. We're product-obsessed, community-powered, and backed by top-tier investors.

We work with thrilling problems like encoding brand voice and visual identity into adaptive systems, designing agentic AI architectures that drive growth, creating tools that let marketers build empathetic flows without code, process automation with AI, etc.

💫 Role
We're hiring a Senior Product Engineer, AI Workflows.

This is not a software engineer who will pick up AI along the way. We want someone who already builds production AI workflows, and who the rest of us can learn from. You should be shipping in weeks, not spending months catching up on how agents work.

We also expect you to own outcomes, not tickets. You talk to users. You decide what to build, with product. You ship it. You're accountable for whether it moved the numbers.

This team sits closest to revenue. You'll work alongside marketing and sales, not at a distance from them.

What you will be doing? 🔧
These are live problems, not hypotheticals.

Quality that survives volume. Generating one good page is a demo. Generating hundreds, all on brand, all schema-ready, is the product. The interesting question is how you know they're good. At that scale, nobody can read them all.

Personalisation that search engines still trust. Every account or segment gets its own variant. Push personalisation too far and you break crawlability and indexing. Push too little and it stops converting. Living in that tension is most of the job.

Closing the feedback loop. Content intelligence tells us what's visible, what ranks, what competitors are doing, and how audiences feel. Feeding that back so the system decides what to build next is largely unbuilt.

Some of these will be solved before you start. Others will replace them. The shape of the work won't change: orchestrating agents in production, integrating external data sources, holding quality steady as volume grows, and staying accountable to numbers a marketer actually cares about.

Are you the one?: 🧠

This role is AI workflow engineering first. You should bring:

• Production experience building AI workflows. Not prototypes: systems that ran under real traffic and failed in interesting ways.

• Hands-on work with LangGraph or LangChain, or an equivalent stack such as the Vercel AI SDK.

• Fluency with model APIs at the level that matters: system and user prompts, tool and function calling, structured outputs, and what to do when the model returns something you didn't expect.

• Strong TypeScript, with solid asynchronous fundamentals. Concurrency, streaming, and error propagation should be second nature.

• Real experience integrating external systems, including CRMs, third-party APIs, and bulk data ingestion, as well as the parts vendors don't document.

Alongside that, you're a capable engineer generally: comfortable across backend and frontend, at home with databases and AWS, and able to make progress on ambiguous problems without waiting to be told what to do.

We weigh competencies over years of experience or job titles.

Bonus points

• Evaluations. If you've designed evals for LLM systems, tell us. It isn't a hard requirement, but it's the area where you'd most obviously teach us something. On this team, judging generated output at volume is the central problem.

• You understand how search actually works: structured data, crawlability, indexing, and how AI answers change the picture.

• You've built for marketing or sales teams and know what they measure.

• You've worked on retrieval systems or vector search.

• You've worked in high-growth SaaS.

Don't tick every box? Apply anyway. We care about how you think, how you build, and how you raise the level of the system around you.

What are the perks? 🎉

- Firstly, we are hiring across Europe

- Latest Macbook

- A budget for you to equip your home office setup

- English classes for all levels

- Online and onsite yoga classes 3x/week

- Yearly company gatherings to take a break from the routine and give you the chance to meet the international teams! (also, other benefits that may depend on the country you're based in)

Work across borders. Come together in Paris 🌍

We’re an international team, with Paris as our hub for collaboration and connection. Every quarter, everyone at Prismic meets in Paris for our In-Office Week. To build together, share ideas and have fun as a team. Attendance is expected, with dates shared well in advance so everyone can plan.

You’ll work from your country of employment, with arrangements depending on your role and location:

• Based in Paris: join us in the office once a week.

• Elsewhere in France: join us in Paris once a month.

• Outside France: work primarily remotely, alongside our quarterly Paris get-togethers.

This role is AI workflow engineering first. You should bring:

• Production experience building AI workflows. Not prototypes: systems that ran under real traffic and failed in interesting ways.

• Hands-on work with LangGraph or LangChain, or an equivalent stack such as the Vercel AI SDK.

• Fluency with model APIs at the level that matters: system and user prompts, tool and function calling, structured outputs, and what to do when the model returns something you didn't expect.

• Strong TypeScript, with solid asynchronous fundamentals. Concurrency, streaming, and error propagation should be second nature.

• Real experience integrating external systems, including CRMs, third-party APIs, and bulk data ingestion, as well as the parts vendors don't document.

Alongside that, you're a capable engineer generally: comfortable across backend and frontend, at home with databases and AWS, and able to make progress on ambiguous problems without waiting to be told what to do.

We weigh competencies over years of experience or job titles.

Bonus points

• Evaluations. If you've designed evals for LLM systems, tell us. It isn't a hard requirement, but it's the area where you'd most obviously teach us something. On this team, judging generated output at volume is the central problem.

• You understand how search actually works: structured data, crawlability, indexing, and how AI answers change the picture.

• You've built for marketing or sales teams and know what they measure.

• You've worked on retrieval systems or vector search.

• You've worked in high-growth SaaS.

Don't tick every box? Apply anyway. We care about how you think, how you build, and how you raise the level of the system around you.