SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Moss is a Finance AI platform for Europe's mid-sized businesses, automating card issuing, invoice management, expenses, and financial workflows. The company has raised €200+ million, operates across Germany, Netherlands, UK, and EU markets, and serves 5,000+ businesses including Flink, Schufa, and Auto1.
You will join as an Applied AI Engineer to build and ship AI agents and intelligent product capabilities that transform how finance teams work. This is a product engineering role—not isolated prototyping or research. You will own agent-based features end to end: from architecture and evaluation through backend integration, deployment, and production operation.
Key responsibilities include:
**Build and Ship AI Agents**: Design and build agents automating complex finance workflows. Own features from concept and prototype through production deployment. Integrate agents deeply with backend services, APIs, data models, permissions, and product workflows. Design secure data access across the Moss platform. Turn emerging AI capabilities into reliable, customer-facing features.
**Evaluate and Improve Agent Performance**: Build systematic evaluations for agent quality, reliability, and business impact. Create representative test datasets, evaluation criteria, regression tests, and human-review processes. Measure and improve accuracy, latency, cost, and user experience. Establish observability and feedback loops for continuous improvement.
**Develop AI Application Architecture**: Apply context-engineering techniques (RAG, MCP, knowledge graphs). Design prompts, tools, memory, workflows, and orchestration strategies. Select appropriate frameworks (Google ADK, LangGraph, LangChain, LlamaIndex). Build guardrails, approval steps, and human-in-the-loop controls for sensitive financial workflows.
**Integrate Machine-Learning Capabilities**: Collaborate with data scientists to integrate ML models into production systems. Build services, data flows, APIs, and operational tooling. Take responsibility for production integration rather than handing off prototypes.
Moss operates 5 offices across Europe with 300+ people from 50+ nationalities. The company offers top-of-market compensation including equity, 20 days "work from abroad," €600 Learning & Development budget, and vibrant office culture with weekly breakfasts and Friday demos.
**Requirements**:
- Seasoned software engineer with experience building and operating production applications
- Highly proficient in Python and/or Java; comfortable with backend services, APIs, data, and application architecture
- Built and shipped at least one agent or LLM-powered product capability end to end
- Personally integrated agents into broader product and backend architecture (not standalone prototypes only)
- Practical experience evaluating agents or other non-deterministic AI systems
- Understanding of providing agents with appropriate data and context while respecting security, permissions, and privacy
- Hands-on experience with prompt engineering, context engineering, and agent orchestration
- Balance rapid experimentation with reliable, maintainable production engineering
- Clear communication and effective collaboration across engineering, product, and data science
- Relevant technologies (not all required): agent orchestration frameworks, context engineering (RAG, MCP, knowledge graphs), AI evaluation methods, language models (Gemini, OpenAI, Anthropic, Llama, Mistral), backend engineering (Python/Java, REST APIs, Kafka, microservices), data systems (SQL, PostgreSQL, BigQuery, vector search), cloud AI platforms (GCP/Vertex AI)
