Django DRFNext.jsTypeScript

Aimyable - AI-Powered Accounts Payable Automation SaaS

An AI-powered accounts-payable automation SaaS built with Django DRF, Next.js, TypeScript, PostgreSQL, Docker, and AWS. The platform combines multi-agent orchestration, invoice OCR and document processing, database automation, RPA-driven UI actions over WebSockets, multi-tenant authentication and RBAC, and self-service onboarding workflows to automate accounts-payable tasks end to end.

The Problem

Coordinating multi-step invoice automation across specialized AI agents, database operations, and RPA-driven UI actions. Extracting structured invoice data from documents so information could flow into the system without manual entry. Supporting secure multi-tenant onboarding, provisioning, role-based access, invitations, and machine-to-machine authentication. Keeping backend services, data models, frontend workflows, and Windows RPA execution synchronized in a production SaaS environment.

The challenge was to create a dependable system that stayed understandable for users while handling complex backend workflows, integrations, and delivery constraints.

The Solution

I broke the system into focused product and engineering layers, then shaped the user experience around clear workflows, resilient infrastructure, and maintainable implementation patterns.

System Architecture

Built a multi-agent workflow engine with specialized Step, DB, and RPA agents for multi-step invoice automation.

Product Execution

Worked on invoice ingestion and data extraction using the Google Cloud Vision Python SDK to OCR invoice documents and parse structured fields.

The Results

The implementation delivered measurable gains across performance, maintainability, and product clarity.

99.9%Reliability Target

40%Workflow Improvement

3xDelivery Velocity