Neural Data Pipeline

Serverless ETL Infrastructure for Marketing Intelligence
7
Lambda Functions
3
Data Collectors
3
Transformers
Parquet
Storage Format
Daily
Orchestration
Data Flow
External APIs
Lambda Collectors
S3 (Parquet)
Orchestrator
Lambda Transformers
Destinations
Core Capabilities
Pipeline Orchestration
parallel
Central orchestrator invokes all transformers in parallel. Consolidated Slack alerts with per-transformer status. Handles failures gracefully without blocking other pipelines.
Lambda Invoke Parallel Execution Slack Alerts
Platform Transformers
3 destinations
Read Parquet from S3, transform to platform-specific formats. GA4 Measurement Protocol, Meta Conversions API, Klaviyo Events API. Server-to-server, no client-side tracking.
GA4 S2S Meta CAPI Klaviyo
Economic Intelligence
2 sources
FRED API for macroeconomic indicators (interest rates, GDP, CPI). Census API for demographic and business data. Enriches marketing data with external context.
FRED API Census API Parquet
Website Analytics
collector
Processes JSON events from S3 (via website Lambda). Outputs events.parquet, conversions.parquet, sessions.parquet. Daily runs at 10:00 AM UTC for previous day's data.
S3 JSON Parquet Output Daily Schedule
Infrastructure as Code
Terraform
Full Terraform modules for Lambda, S3, EventBridge. Environment-specific configs (dev/prod). Lifecycle rules for data retention and S3 Glacier transitions.
Terraform EventBridge Modular IaC
Monitoring & Alerts
integrated
Sentry for error tracking across all Lambdas. Slack webhooks for pipeline status. CloudWatch logs with structured JSON. Per-transformer success/fail counts.
Sentry Slack CloudWatch
AWS
Lambda (7) S3 EventBridge Secrets Manager CloudWatch IAM
Stack
Node.js 20 TypeScript (strict) parquetjs Terraform
Schedule
Website Analytics: 10:00 UTC FRED: 10:30 UTC Census: Quarterly Orchestrator: 10:45 UTC
Outputs
GA4 Measurement Protocol Meta Conversions API Klaviyo Events API Neural Core Data Lake