Back to projects

MARIA - AI-Powered Government Transparency Platform

MARIA - AI-Powered Government Transparency Platform
Amrood Labs

Project Overview

MARIA is an AI-powered transparency platform processing Dominican Republic government documents, serving real-time data to frontend applications and external integrations.

Developed REST APIs for MARIA platform using Node.js and Express with MongoDB, implementing scalable microservices API architecture.

Built scalable microservices API architecture implementing Redis-based job queues (BullMQ) for concurrent document processing workflows.

Implemented AWS S3 API integration for secure document storage with presigned URL generation and CloudFront API for optimized content delivery to client applications.

Designed event-driven API system with separate service endpoints for PDF processing, image conversion, and AI analysis, ensuring fault-tolerant document processing pipelines.

Deployed production APIs using PM2 with automated service orchestration, health check endpoints, and monitoring APIs for system reliability and performance metrics.

Key Features

REST API Architecture

Developed REST APIs serving real-time data to frontend applications and external integrations with scalable microservices architecture.

Document Processing Workflows

Built Redis-based job queues (BullMQ) for concurrent document processing workflows ensuring reliable and efficient processing.

AWS Integration

Implemented AWS S3 API integration for secure document storage with presigned URL generation and CloudFront API for optimized content delivery.

Event-driven System

Designed event-driven API system with separate service endpoints for PDF processing, image conversion, and AI analysis.

Production Deployment

Deployed production APIs using PM2 with automated service orchestration, health check endpoints, and monitoring APIs.

Challenges & Solutions

Concurrent Document Processing

Processing multiple government documents concurrently while maintaining system stability and performance was challenging.

Solution:

Implemented Redis-based job queues (BullMQ) for concurrent document processing workflows, ensuring reliable and scalable processing pipelines.

Secure Document Storage

Ensuring secure storage and optimized delivery of government documents required robust cloud integration.

Solution:

Implemented AWS S3 API integration for secure document storage with presigned URL generation and CloudFront API for optimized content delivery.

Fault-tolerant Processing

Building fault-tolerant document processing pipelines with separate service endpoints was complex.

Solution:

Designed event-driven API system with separate service endpoints for PDF processing, image conversion, and AI analysis with health check endpoints.

Project Gallery

MARIA platform processing government documents

MARIA platform processing government documents

Project Details

Timeline

2023 - Present

Company

Amrood Labs

My Role

Backend Developer

Technologies Used

Node.js
Express
MongoDB
Redis
BullMQ
AWS S3
AWS CloudFront
PM2
REST APIs
Microservices
Event-driven Architecture