AI Engineer [Feelance]
About the team
At Equativ, we're on a mission to develop advertising technologies that empower our customers to reach their digital business goals. The impact of Generative AI on the industry is projected to be major, and Equativ has been undergoing a significant transformation to embed this technology at the core of our value proposition.
The GenAI team is a new, strategic unit (6–8 people) responsible for the design, development, and deployment of Equativ's entire GenAI stack. It is split into two complementary pillars:
-
Core Platform — designing and building a reusable, industrialized Agentic Platform (MaaS, AaaS — Model/Agent as a Service) to accelerate agent development across the company.
-
Business Feature — the team you'll be joining. We develop and deploy mission-critical AI agents that automate workflows for non-technical internal teams (operations, ad quality, marketplace, finance support, etc.). Our backlog targets people who today rely on manual processes and ad-hoc tooling — your work will give them autonomous, measurable, production-grade AI assistants.
This team is part of the R&D department (200+ engineers across Paris, Nantes, Limoges, Krakow, Berlin and North America), all working in an Agile environment.
What you will do
-
Design, develop, and deploy goal-oriented AI agents and full agentic systems using Google ADK and n8n, automating high-impact workflows inside operational teams.
-
Partner directly with non-technical stakeholders — sit with operations teams, understand their pain points, translate workflows into agentic solutions, and explain your design choices in language they actually understand. This is roughly half the job.
-
Be the team's Python referent. You will own the production codebase, enforce strong software engineering principles (unit tests, CI/CD, Git, code review), and bring AIOps/AgentOps best practices to the team (observability, versioning of agent blueprints, eval pipelines).
-
Ship complete agentic systems end-to-end — orchestration, evaluation, guardrails, prompt management, and integration with internal data sources via MCP servers that you'll build yourself when needed, or partner with other R&D teams when they own the system. If a partner team doesn't have bandwidth, you should be comfortable picking up the integration work directly.
-
Detect platform-level gaps (reliability, latency, observability, model routing) and formulate clear, prioritized requests to the GenAI Core Platform team. You don't own the platform — but your production usage is what pushes it to get better.
-
Build a working expertise in agentic design patterns (eval, guardrails, multi-agent orchestration) and share it with internal AI champions and AI builders as they emerge across the company.