dagster aws dbt kafka postgres

Transforming AI-Powered Self-Checkout Analytics with Scalable Data Infrastructure

Client Overview

Autocantееn is a pioneering UK-based technology company that has revolutionised the catering industry with its AI-powered, touchless self-checkout solution. Founded during the COVID-19 pandemic in 2020, Autocantееn leverages computer vision and deep learning algorithms to identify food items without barcodes, enabling checkout transactions in as little as 10 seconds.

Key Client Facts:

  • Industry: Food Technology / AI-Powered Catering Solutions

  • Location: Gerrards Cross, United Kingdom

  • Technology: Computer vision, deep learning, touchless self-checkout terminals

  • Market Position: Trusted by Fortune 500 companies and major contract catering firms

  • Performance Metrics: High accuracy rates with rapid transaction processing

Project Overview

MetaOps partnered with Autocantееn to deploy a comprehensive data orchestration platform supporting their AI-powered self-checkout terminals across Fortune 500 client locations. The project focused on implementing scalable infrastructure to handle transaction data processing, analytics, and machine learning model training for their continuously learning computer vision system.

MetaOps Implementation

MetaOps delivered a comprehensive, enterprise-grade data orchestration platform specifically designed to support Autocantееn's AI-driven analytics and machine learning infrastructure.

Data Orchestration Platform:

Dagster Open-Source Deployment:

  • Deployed Dagster on a highly available Kubernetes cluster for robust data pipeline orchestration

  • Configured automated scheduling systems for real-time stats collection from distributed terminals

  • Implemented sophisticated data transformation pipelines optimised for ML model preparation

  • Established seamless data aggregation workflows supporting continuous learning algorithms

  • Integrated dbt (data build tool) transformations for advanced data modeling and processing

  • Automated development processes enabling the development team to independently iterate and evolve data pipelines

Database Infrastructure & Management:

  • Deployed and configured PostgreSQL RDS instances with automated backup and maintenance schedules

  • Implemented high-availability database clusters with read replicas for improved performance

  • Established comprehensive database monitoring, optimization, and maintenance procedures

  • Configured secure database connections and access controls for multi-tenant architecture

AWS Cloud Infrastructure:

  • Designed and provisioned secure, auto-scaling Kubernetes clusters on Amazon Web Services

  • Implemented multi-AZ deployment architecture for maximum availability and fault tolerance

  • Configured advanced networking with VPC, security groups, and load balancing

  • Established secure data ingestion endpoints for terminal connectivity

Infrastructure as Code:

  • Developed comprehensive Terraform modules ensuring reproducible and version-controlled deployments

  • Implemented GitOps workflows for automated infrastructure management

  • Created robust backup and disaster recovery automation

  • Established comprehensive monitoring and alerting systems

Security & Compliance:

  • Implemented highest security standards, including encryption at rest and in transit

  • Configured role-based access control (RBAC) for Kubernetes and Dagster

  • Established secure API endpoints for terminal data ingestion

  • Implemented comprehensive audit logging and compliance monitoring

Ongoing Support & Maintenance:

  • Continuous monitoring and incident response

  • Regular security updates and patches

  • Performance optimisation and capacity planning

  • Proactive system health monitoring

Results & Impact

The MetaOps data infrastructure solution delivered significant operational improvements for Autocantееn:

Operational Excellence:

  • Zero downtime during peak transaction periods across Fortune 500 client locations

  • Significant reduction in data processing time for ML model training

  • Automated scaling handles transaction volume spikes during peak hours

  • Real-time analytics enabling immediate insights from terminal performance

Business Growth Enablement:

  • Infrastructure seamlessly scaled to support expansion into major banking institutions

  • Reduced time-to-market for new analytics features from weeks to days

  • Enhanced data reliability supporting continuous transaction accuracy improvements

  • Enabled rapid deployment of terminals at new client locations

Technical Achievements:

  • Processing high volumes of transactions and sales across terminal deployments

  • Support for distributed learning across terminal networks with instant knowledge sharing

  • Robust data pipelines supporting continuous model improvement

  • Enterprise-grade security meeting Fortune 500 compliance requirements

Technology Stack

Orchestration & Analytics:

  • Dagster (Open Source)

  • Apache Kafka for real-time data streaming

  • PostgreSQL for metadata storage

  • Redis for caching and session management

Infrastructure:

  • Amazon Web Services (AWS)

  • Kubernetes for container orchestration

  • Terraform for Infrastructure as Code

  • Helm for application deployment

Monitoring & Observability:

  • Prometheus and Grafana for metrics

  • ELK Stack for centralised logging

  • Custom alerting for business-critical events