AI-Powered Accounting & Financial Intelligence Module (IFRS-Compliant, Multi-Entity, Project-Based) Project Overview We are building an advanced internal financial system on top of an existing FastAPI + PostgreSQL backend. We are looking for a senior backend engineer or small experienced team to design and implement a structured, AI-assisted accounting engine with OCR, reconciliation, and full financial statement generation. This is not a basic bookkeeping tool. It is a structured financial intelligence module integrated into a broader AI architecture. Budget & Timeline Delivery Deadline: 6 weeks from project start Milestone-based payments Serious candidates only. Architecture quality matters. Scope of Work 1. OCR & Document Processing Ingest PDFs, images, bank statements, invoices. Extract structured financial data (date, counterparty, amount, tax info, references). Store metadata and link to accounting entries. Flag low-confidence extractions. Support English and Spanish documents. 2. Automated Double-Entry Accounting Engine Full double-entry bookkeeping logic. Versioned rule-based classification system. Automatic journal entry generation. Manual override with audit trail. Prevent unbalanced entries. Allow recalculation if rules are updated. Core components: Chart of Accounts (aligned with IFRS) Bank transactions Journal entries and lines Accounting rules (versioned) Audit logs 3. IFRS-Compliant Financial Statements The system must generate: Balance Sheet Income Statement Cash Flow Statement (Direct and Indirect) Statement of Changes in Equity Must support: Multi-entity Multi-project Multi-currency Monthly and annual reporting Reproducible financial snapshots 4. Project-Based Accounting Every transaction must be allocatable to a project. Project-level P&L and cash flow. Revenue recognition logic configurable. AI-assisted classification when ambiguous. 5. Related Parties & Intercompany Logic Structured handling of related-party transactions. Due-to / due-from entries. Internal service billing logic. Cross-entity cost allocations. Intercompany reconciliation. No personal data hardcoded — fully entity-based structure. 6. Reconciliation Engine Bank reconciliation Invoice reconciliation Intercompany reconciliation Adjustment proposals Full reconciliation history 7. AI Integration AI must assist with: OCR validation Transaction classification fallback Anomaly detection Pattern recognition AI must not replace accounting logic. All AI outputs must be logged and reviewable. Technical Requirements Python (FastAPI compatible) PostgreSQL Clean database architecture Alembic migrations Modular design API-first structure Future integration readiness (CRM, BI, banking APIs) Architecture must be clean, extensible, and integration-friendly. Deliverables Full database schema Accounting engine implementation OCR integration Rule engine Reconciliation module Financial statement generator AI integration layer Documentation Deployment guide Basic test coverage