Software Engineer - AI

ABOUT THE ROLE This is a full-stack AI software engineering role on a 15-person product and engineering team building AI-powered tools for construction professionals. You will own end-to-end AI features from concept through production while also contributing to the platform foundations that make new use cases fast to ship. Your work will shape both customer-facing experiences and the infrastructure beneath them. WHAT YOU'LL DO - Build and own end-to-end AI features including document versioning and agentic chat workflows, from database schema through to UI. - Manage evaluations, scoring, and error analysis on new and shipped AI features to maintain quality in production. - Contribute to the workflow harness, decision-trace standard, and evaluation tooling that accelerates shipping new use cases. - Build reusable human review and approval surfaces that build trust in AI output. - Model and evolve a construction domain ontology and project graph, including entity resolution across documents such as bills of quantities, contracts, and drawings. - Design features that link answers to evidence and make decisions auditable and traceable. - Extend multi-modal ingestion of unstructured documents including PDFs, scans, drawings, and 3D models at scale. - Build read/write integrations with customer systems such as document management platforms and ERPs. - Contribute to scalable data pipelines and support native BIM and DWG file processing. WHAT WE'RE LOOKING FOR - 3 or more years of experience as a full-stack software engineer delivering production-grade applied AI features to customers. - Proficiency with TypeScript, Python, and SQL for building full-stack applications. - Experience building backend APIs and database schemas using frameworks such as Fastify, Prisma, or equivalents. - Experience integrating and optimizing LLM APIs (OpenAI, Azure, Gemini, Cohere, or similar). - Experience with cloud infrastructure such as AWS (RDS, S3, Elastic Beanstalk) or equivalents. - Experience building data pipelines and ETL workflows for processing unstructured data at scale. - Experience with vector databases and semantic search implementations. - Experience designing and implementing evaluation frameworks and error analysis for AI systems. - A track record of building internal platform or developer-facing tooling that other engineers ship on top of. - B2B product experience in a startup or early-stage company environment. - Fluent spoken and written English; German is a bonus but not required. - Nice to have: experience with workflow orchestration tools (Prefect, Airflow, or similar), document processing or OCR, and observability tools (Datadog, Langfuse, or similar). COMPENSATION & BENEFITS Salary range: 80,000 to 120,000 EUR annually. Visa sponsorship is not available for this role. LOCATION On-site, five days per week in Berlin, Germany. Candidates must be based in Berlin or willing to relocate.