Dataleon

Sr. Full Stack Developer

AI automation platform with ML model integration and data extraction

2022•8 months
ReactNode.jsElectron.jsServerlessAWS CognitoAWS RDSAWS LambdaAWS Textract
ai
D
dataleon

Dataleon

AI automation platform with ML model integration and data extraction

ReactNode.jsElectron.jsServerlessAWS Cognito
Explore →

Problem Solved

Automated invoice and document data entry by combining AWS Textract OCR with ML classifiers — replacing manual keying that was error-prone and slow with a fire-and-forget event pipeline.

Overview

  • •Integrated ML models for automated data processing
  • •Implemented AWS Textract for image-based data extraction
  • •Built serverless architecture with AWS Lambda and event-driven patterns
  • •Developed admin management systems
  • •Integrated Google Places API for location services
  • •Built Electron.js desktop application variant

Key Skills Demonstrated

AWS Textract OCR IntegrationML Model DeploymentEvent-Driven Serverless Architecture (SNS/SQS)Electron Desktop App DevelopmentGoogle Places API IntegrationAWS Cognito IdentityLambda + RDS Patterns

Challenges

  • •Integrating ML models for production use
  • •Extracting data accurately from images
  • •Building event-driven serverless architecture

Solutions

  • •Implemented AWS Textract for OCR and data extraction
  • •Built serverless pipeline with Lambda, SNS, and SQS
  • •Created Electron desktop app for offline capabilities

Key Achievements

  • Cut manual data-entry time substantially via automated OCR + classification
  • Built event-driven SNS/SQS pipeline that handled bursty document loads
  • Shipped a parallel Electron desktop app for offline use cases
  • Integrated Google Places for location-aware data enrichment

Tech Stack

Frontend

React.jsElectron.js

Backend

Node.jsServerlessAWS LambdaCognitoRDSTextractSNSSQS

Deployment

Deployed on AWS with Lambda, RDS, and event-driven architecture using SNS and SQS

Third-Party Integrations

AWS TextractGoogle Places APIAWS SNSAWS SQS