Forward Deployed AI Engineer - AI Agent Ops

At Smalk, around 40 AI agents already run parts of the business. We’re past the “should we try AI?” phase. Now we need someone whose full-time job is making those agents genuinely good, and then taking that know-how to our clients.

If you’ve ever built a custom GPT, a Claude skill or a small MCP server that your team treated as a side project: here, that’s the job.

The future we’re building

Advertising was built for browsers. The future belongs to agents. More than 800 million people now use AI search every week, and the answers they get from ChatGPT, Perplexity, Claude or Google AI Mode draw on publisher content.

Smalk is building the economy of that new world. We help publishers thrive in the AI Search era with a brand-new revenue stream, and help brands earn their place at the top of AI answers. That’s the category we’re pioneering: Generative Engine Advertising (GEA).

Your first 12 months

Months 0 to 3: learn the company through its workflows. You sit with sales, ops, marketing, product and tech, then ship your first skills and connectors for them. You take over our business agents and learn what breaks.

Months 3 to 6: own internal AI end to end. A team brings you a need and leaves with something they use without you. You measure what’s used, fix what isn’t, and document everything.

From month 6: go client-facing. You help agencies and brands plug Smalk’s MCP server into their own stack, help them write skills that use Smalk data, and improve the AI help agents on our platform.

What you’ll own

• Skills, commands and workflows: reusable Claude and OpenAI skills, slash commands and prompt workflows for every team.
• Connectors: MCP servers and integrations (Notion, Slack, Google Workspace, Linear, CRM) so agents work with real company data.
• Our business agents: keeping our ~40 in-house agents reliable and improving their prompts, coverage and robustness over time.
• Client deployments: hands-on integration of Smalk’s MCP into client tools and workflows.
• Platform help agents: prompts, knowledge, evaluation and edge cases.
• The feedback loop: bringing what you learn in the field back to product and tech.

You’ll report directly to Antonin, our CTO and co-founder, and Benjamin, our founding Product Manager.

Tech stack

Claude (skills, MCP, Claude Code), OpenAI (Agents, connectors, Apps SDK), MCP servers. Python, Django, DRF, PostgreSQL, Redis and Celery on our side. Notion, Slack, Google Workspace and Linear internally.

How we work

We believe developers are AI Agent Team Leads, not craftsmen typing code line by line. This role is the purest version of that: your job is to make every person at Smalk, and then every client, 10x more effective with AI. What matters is your curiosity, your ability to understand a business need, and how fast you turn it into something that works.

You’ll probably fit if you’ve already

• Built a GPT, skill, MCP server or agent that someone other than you actually used.
• Made an LLM reliable rather than just impressive in a demo, and can explain how.
• Written enough Python to call an API and handle the JSON that comes back.
• Explained AI to someone who didn’t want to hear about AI.
• Broken things while tinkering, and fixed them.

You’ll need fluent French and working English.

BTS, coding bootcamp, engineering school, self-taught: we care about what you’ve built, not where you studied.

Bonus points if you’ve published your own MCP server or skill, automated workflows for a team, worked with APIs, webhooks or OAuth, or you’re curious about adtech.

What we offer

• Permanent contract (CDI).
• €45k to €55k gross per year depending on profile.
• Paris (Station F), Bordeaux, hybrid or full remote, with regular team touchpoints.
• Direct access to the CTO and fast ownership.
• A seat right next to our biggest customers, helping them in their day-to-day work.
• Pioneering a new category and being ahead of your time.
• Joining an AI-native company in the front row of AI Search.

Hiring process

We’ll ask you to show us something you’ve built with AI, then to tackle a hands-on case: designing a skill or connector for one of our teams.