Data Infrastructure Summary

Effective Date: January 1, 2026 Version: 1.1 Last Updated: March 1, 2026

This document provides a technical summary of the infrastructure that stores, processes, and transmits customer data within the Estimatics Platform. It is intended for enterprise customers, compliance teams, and technical stakeholders conducting due diligence.

1. Data Classification

Data Type	Description	Storage Location
Account Data	User profiles, credentials, organization settings	Aurora PostgreSQL (RDS)
Job Metadata	Job records, statuses, assignments, timestamps	Aurora PostgreSQL (RDS)
Media Files	Photographs, videos, LiDAR scan data	Amazon S3
Processed Outputs	AI findings, scopes, reports, geometry	Aurora PostgreSQL + S3
Session Data	Authentication tokens, session state	ElastiCache Valkey
Queue Messages	Background job payloads	Amazon SQS
Logs	Application and access logs	CloudWatch Logs

2. Primary Data Store

Service: Amazon Aurora PostgreSQL 17.4 Region: US East-1 (Northern Virginia) Configuration:

Multi-AZ deployment with automatic failover
Automated daily backups retained for 7 days
Point-in-time recovery within the 7-day window
Encryption at rest using AWS KMS
Private VPC subnet — no public endpoint

3. Media Storage

Service: Amazon S3 Region: US East-1

All customer photographs, videos, and generated documents (PDF reports, export files) are stored in private S3 buckets. Key configurations:

Bucket-level encryption using SSE-S3 (AES-256)
Versioning enabled on production buckets
Lifecycle policies for cost optimization (infrequent access after 90 days)
No public access — all access via pre-signed URLs with expiration
Separate buckets for production and UAT environments

4. Compute and Application Layer

Service: Amazon EC2 with Auto Scaling Groups Configuration:

Minimum 2 always-running instances
Horizontal scaling triggered by load metrics
Launch Templates with pre-configured AMI for fast instance initialization
Elastic File System (EFS) mounted at /home/www — shared across instances for code and configuration
Application Load Balancer distributes traffic across instances

5. Background Processing

Service: Amazon SQS (Standard Queues) Usage: All background jobs, including AI analysis, media processing, and report generation, are dispatched via SQS. This decouples user-facing requests from compute-intensive processing.

Worker Configuration: Laravel queue workers running under Supervisor on EC2 instances.

6. Caching and Real-Time

Cache: Amazon ElastiCache (Valkey Serverless — Redis-compatible) Endpoint: TLS-encrypted (rediss:// scheme) Usage: Session data, rate limiting, frequently-accessed configuration

WebSocket: Laravel Reverb, managed via Supervisor Scaling: Reverb uses ElastiCache Valkey as a pub/sub backend to support multi-instance WebSocket broadcasting

7. AI Processing Infrastructure

AI tasks are executed via API calls to external AI providers. No AI model training or inference occurs on Estimatics-owned compute infrastructure.

Provider	Data Sent	Retention by Provider
OpenAI	Images, structured text	Not retained (enterprise agreement)
Anthropic	Text, structured data	Not retained (enterprise agreement)

All API calls are made over HTTPS. No raw customer data is stored in AI provider systems beyond the processing request.

8. Data Residency

All primary data storage and processing occurs in AWS US East-1 (Northern Virginia). AI API calls may be processed on provider infrastructure that spans multiple US data centers. No customer data is intentionally stored outside the United States.

9. Backup and Recovery

Component	Backup Frequency	Retention	RTO Target
Aurora PostgreSQL	Continuous (PITR)	7 days	< 4 hours
S3 Media	Versioning (on change)	90 days (lifecycle)	< 1 hour
Application Configuration	EFS snapshots	14 days	< 2 hours

Contact

Technical infrastructure questions: security@aiestimatics.com