Senior Full-Stack Developer / AI Enginee

Job Description We are looking for an experienced full-stack developer or development team to build a modern accounting automation platform for tax advisory firms. The software will help tax firms process large volumes of accounting documents automatically. Users should be able to upload PDFs, images, ZIP files and complete folder structures. The system should analyze the documents, classify them, extract relevant accounting data, match them with bank transactions, create booking suggestions and prepare the data for review and export. The goal is not just a simple document upload tool. We want to build a professional accounting automation system for tax firms, with a strong focus on document processing, AI-assisted data extraction, bank matching, audit logs, security and a clean review workflow. The system should work without n8n or external workflow automation tools. All workflows should be handled internally through backend logic, queues, workers and scheduled jobs. Main Features Required Multi-tenant platform for tax firms and their clients Client portal for document upload Upload of PDFs, images, ZIP files and folder structures Automatic document classification OCR and AI-based document reading Extraction of invoice data such as supplier, invoice number, date, net amount, VAT, gross amount, IBAN and line items Support for German accounting workflows Support for SKR03 / SKR04 account suggestions Automatic booking suggestions Bank statement import and bank transaction matching Review dashboard for accountants and tax advisors Confidence scores for AI results Manual correction and approval workflow Audit log for all important actions DATEV-compatible CSV export Secure file storage User roles and permissions Strict client data separation Admin dashboard for tax firms Optional hybrid / on-premise deployment model Preferred Tech Stack We are open to suggestions, but the preferred stack is: Frontend: Next.js / React Backend: Node.js / NestJS or Python / FastAPI Database: PostgreSQL Queue System: Redis / BullMQ, RabbitMQ or similar File Storage: S3-compatible storage / MinIO AI/OCR: Mistral OCR, OpenAI, local OCR or other document AI models Deployment: Docker / Docker Compose Security: JWT, role-based access control, encryption, audit logs Important Requirements No n8n No low-code workflow tool Clean and scalable architecture Secure handling of sensitive accounting data Code must be well-structured and documented The platform must support large document volumes AI should assist with extraction and suggestions, but all accounting results must be reviewable by a human user The system must be built in a way that future modules can be added easily First Project Scope / MVP The first version should include: User login and role management Tax firm dashboard Client management Client portal Document upload ZIP and folder processing OCR / AI document extraction Document classification Booking suggestion logic Bank transaction import Basic bank matching Review and approval screen Audit log DATEV CSV export Docker-based deployment What We Are Looking For We are looking for someone who can think like a product architect, not just a coder. You should have experience with: SaaS or enterprise software Accounting, finance, ERP or document automation systems AI / OCR document processing Multi-tenant systems Secure backend development API design Docker deployment Clean database design Experience with German accounting, DATEV, SKR03/SKR04, XRechnung or ZUGFeRD would be a strong advantage. Deliverables Complete source code Frontend application Backend API Database schema Worker / queue system OCR / AI integration Docker setup Technical documentation Installation guide Basic test data Admin and user workflow documentation Questions for Applicants Please answer these questions in your application: Have you built accounting, ERP, finance or document automation software before? Have you worked with OCR or AI document extraction? What tech stack would you recommend for this project and why? How would you structure the system without using n8n? How would you handle security and client data separation? Can you build this as a scalable MVP first and then extend it step by step? Please share relevant examples of previous work. NDA / Confidentiality This project involves a serious software product in the accounting and tax advisory market. An NDA may be required before sharing detailed specifications, designs or business logic. Preferred Working Style We prefer milestone-based development with clear deliverables. The first milestone should be a technical architecture plan and clickable dashboard structure before full development starts.