[SKALEET]
Skaleet builds the core banking engine behind banks and neobanks: accounts, payments, cards. Our platform runs in production for 30+ regulated financial institutions in 15 countries, including Crédit Agricole, Caisse d'Epargne and Nirio (FDJ United), and manages over 10 million accounts.
In 2026, Euromoney named Skaleet the best banking solution in Europe. Our revenue grew by 40% in 2025. We are about 100 people, based in Boulogne-Billancourt.
Our ambition: become the platform of choice for European banks and fintechs that want to replace their legacy core banking systems.
Banks also use Skaleet to launch new activities and countries, as Crédit Agricole did in Germany. Our data is hosted in France and Germany: built in Europe, for Europe.
[Job Description]
💡 Context
Skaleet recently created its AI Team, with one clear ambition: improve our clients' banking operations through AI. Running a bank means thousands of back-office operations every day; our goal is to make them faster, more reliable and more scalable.
As an AI Engineer, you will join the AI Team and report directly to Thomas Linck, VP Platform. You will build and own Skaleet's AI products: a suite of specialized banking agents that let back-office operators understand, analyze and act on the platform in natural language, plus AI-powered automations that take over recurring operational processes end to end.
You will ship AI into production banking environments, where "mostly right" is not an acceptable bar.
👉 Your future missions
Architecture & delivery
Build and ship Skaleet's AI products end to end, from early prototypes to production deployment
Own the technical architecture of our agentic platform and continuously improve it by integrating the latest advances in foundation models, agent frameworks and AI engineering
Agents & reliability
Design reliable agentic workflows combining LLMs, tools, banking data and human oversight
Define how agents reason, retrieve context, call tools, handle failures and collaborate to complete complex banking operations
Own AI quality, observability and evaluation, using traces, user feedback, annotated datasets, LLM-as-judge, calibration against human labels and regression gates
Integration & data
Build the integration layer between agents and Skaleet's platform: MCP servers and APIs, production-grade authentication, permission management, error handling, retries, idempotency and structured outputs
Develop the context and retrieval systems powering our agents: embeddings, vector search, hybrid retrieval, reranking and knowledge ingestion over Skaleet's product documentation and clients' business processes
Work closely with data engineers to make banking data usable by AI systems: data models, pipelines, interfaces, tenant isolation
Security & compliance
Implement the safeguards required for banking use cases: human validation gates, permission scoping, tenant isolation, audit trails, explainability, safe failure modes
Product & iteration
Prototype quickly, iterate with banking operators, identify high-value use cases and improve the product based on real operational feedback
Evaluate and integrate the best available AI models and technologies based on quality, latency, cost, security and regulatory constraints
This role is about building systems on top of existing foundation models, not training models from scratch.
[Desired Profile]
✨ You are a great match if:
Must-have
Minimum 5 years of software engineering experience, including significant hands-on experience building LLM-based agent systems in production
Strong Python backend skills: REST APIs, async, automated testing
Concrete experience with tool calling, RAG, agent context/memory management, and LLM quality evaluation
Comfortable designing and integrating complex, data-intensive production systems
Experience monitoring and optimizing LLM systems in production: observability, tracing, cost and latency control at scale
Working knowledge of guardrails and defenses…