Software Backend Engineer, SaaS product (Python) - Greenstep Espoo/Vantaa

Tiivistelmä työpaikkailmoituksesta: Build and own backend services, REST APIs and agentic workflows in Python/FastAPI using LangChain/LangGraph and expose tools via MCP. You must have 3+ years of Python backend experience with async programming, REST API design, ORMs/migrations (SQLAlchemy, Alembic), testing discipline, Docker and practical AI agent-building skills. You will work end-to-end with the data and frontend teams on scalability and the AI/MCP roadmap in a hybrid role based in Espoo or Vantaa, and Greenstep does not sponsor work visas. Greenstep offers a salary model that matches your responsibilities and experience, a 10% annual profit Together bonus, occupational health support and development via Greenstep Academy. Tämä tiivistelmä on luotu tekoälyn avulla. Greenstep is a growing, global group helping thousands of companies in Europe to grow through clarity. We are a technology-oriented company with Finnish roots, founded in 2010. In addition to Finland, we currently operate in Sweden, Norway, Estonia, Denmark, and the Netherlands. Our team of around 800 experts consists of seasoned professionals in areas such as accounting, CFO services, financial management, HR, analytics, ERP and CRM technologies, payroll, as well as tax and legal services. We are an agile growth company, and our service offering continuously expands in line with our customers’ evolving needs. The Role Our platform consolidates client operative business data into a single place and turns it into real product capability — BI on Power BI, integrations, AI/Copilot, MCP, and features built directly on top of semantic models and gold-layer data, not just raw access to it. As a Backend Engineer, you'll build the AI and agent layer that makes that possible: designing backend services and APIs in Python/FastAPI, building and orchestrating AI agents with LangChain/LangGraph, and exposing them through MCP — a genuine strategic bet for us, not a checkbox. You'll own the service, API, and agent layer end-to-end, from design through deployment , working closely with our data team (who own the medallion architecture and warehouse modeling) and a dedicated frontend team. You'll sit close to the product owner driving our scalability and AI/MCP roadmap, with real influence over how agentic features get architected — this is a build role, not a maintain-what-exists role. You will : Design and develop backend services and REST APIs in Python/FastAPI, including agentic workflows built with LangChain and LangGraph. Build and expose tools/agents via MCP (FastMCP). Build features and APIs on top of consolidated semantic-model and gold-layer data — turning data into product, not just serving it up raw. Design database schemas and migrations (SQLAlchemy & Alembic). Write asynchronous, production-grade Python, with solid test coverage including mocking LLM/agent calls. Containerize services with Docker for local dev and deployment — you won't own Kubernetes or infra, but your services need to ship clean. Instrument agent behavior with real observability, so non-deterministic failures are debuggable, not mysterious. Use AI coding tools to move faster, with engineering judgment intact — we want a builder who uses AI well, not one replaced by it. What we're looking for You bring 3+ years of Python backend experience and genuine interest in AI agent systems. This isn't a role for someone who just wants to consume an LLM API and call it done. You reason well about async and distributed workflows, take ownership of your work, write clean and well-tested code , and communicate clearly, especially when explaining agent behavior that isn't always deterministic. We work mainly in Finnish . Required skills: 3+ years professional Python backend development, including async programming. OOP and clean service architecture . REST API design ( FastAPI ). ORMs & migrations (SQLAlchemy, Alembic). Practical AI agent building — agent logic, tool use, state (LangChain, LangGraph). Testing discipline : unit/integration/API tests, mocking LLM calls. Docker (containerization — no orchestration ownership needed). Effective use of AI coding tools without losing engineering judgment. Basic CI familiarity. Excellent English , spoken and written. Finnish is preferred. Preferred: FastMCP / MCP server-building experience. Power BI, Microsoft Fabric, or BI/analytics platform experience. Machine learning knowledge; vector search, embeddings, RAG. dbt or Databricks familiarity. Task queues (Celery, arq) and Redis for agent workflows. Experience building features on top of semantic layers/warehouses, or integration platforms consolidating multiple source systems. Why Greenstep? Greenstep is a family-owned company founded in 2010 with no loans and a strong balance sheet. Since 2010, we have grown to almost 800 people, more than 3000 customers, and we are looking forward to continuing this growth journey! We provide the full spectrum of financial services , from accounting to int