Staff Software Engineer, AI Agents (Python)

Staff Applied AI Engineer, Backend Join us as a Staff Applied AI Engineer, Backend to build production systems in which AI is a core runtime capability. You'll help Qonto's Anti-Financial Crime teams investigate cases faster and with greater confidence by turning complex operational workflows into dependable, measurable tools. As a Staff Applied AI Engineer you will Build production AI systems: Design and ship agentic tools and AI-powered workflows that gather context from multiple internal systems, orchestrate models and tools, parse structured outputs, and handle uncertainty and failure safely Own projects end to end: Lead discovery with AFC stakeholders, shape the solution, make architectural decisions, implement and launch it, then operate, maintain, and improve it in production Design robust backend foundations: Build reliable, maintainable, and extensible services, APIs, databases, and integrations around evolving AI models and tooling Make AI behaviour measurable: Define evaluation approaches, observability, fallbacks, and human-review mechanisms; monitor output quality, acceptance and edit rates, throughput, and operational impact Deliver measurable operational value: Reduce investigation lead time and expand automation across AFC workflows through pragmatic, incremental delivery Shape the team's technical direction: Lead design discussions, anticipate risks, balance speed with quality, and help raise the team's capability in production agentic systems What you can expectAI at the heart of the system: This is not a conventional backend role using AI only as a coding assistant, nor an ML research role. You'll build real products where model behaviour, orchestration, evaluation, and failure handling are production concerns High autonomy: There is no dedicated Product Manager. Engineers work directly with AFC stakeholders and own the path from an ambiguous operational need to a measurable production outcome Lean, iterative delivery: The team uses a daily 15-minute blocker sync, bi-weekly 1:1s, and lightweight tracking, leaving engineers focused on building and solving problems A close user feedback loop: You'll collaborate directly with operational teams and measure success through investigation lead-time reduction, output quality, human acceptance and edit rates, adoption, throughput, and resources saved A complex, meaningful domain: You'll learn how to build safe, scalable AI automation in a regulated environment where reliability and auditability matter