Skip to content
RelentlessData Engineering

Work

Real engagements, mapped to the services we run: platform builds, ingestion at scale, activation and governance. Each case study includes the challenge, the architecture, the trade-offs and the measured outcomes.

  1. 01

    Multi-tenant ingestion at scale: re-architecting CDC for ~2,000 tenant schemas

    Sector
    B2B SaaS
    Duration
    12 months

    How a multi-tenant SaaS platform with ~2,000 PostgreSQL tenant schemas re-architected ingestion on AWS DMS, S3 and Snowpipe, and consolidated a union-of-everything schema sprawl into one table per entity. That cut ingestion cost by 90% and transformation cost and latency by 80%, took about $500,000 a year out of compute spend and brought source-to-queryable latency down to about 3 minutes.

    • $500,000/yearsaved in compute cost (indicative)
    • 90% loweringestion cost (indicative)
    • 80% lowertransformation cost and latency (indicative)

    Stack

    • PostgreSQL
    • AWS DMS
    • Amazon S3
    • Snowpipe
    • Snowflake
    • Python
    • dbt
    • GitHub Actions
    • Docker

  2. 02

    30 days to data-driven: a governed analytics platform with Snowflake, AWS, GitLab CI/CD, dbt, Fivetran, Census, Metabase and Power BI

    Sector
    B2B SaaS
    Duration
    30 days

    How we stood up a governed analytics platform on Snowflake, AWS, GitLab CI/CD, dbt, Fivetran, Census, Metabase and Power BI in 30 days, giving the client 80% faster insights, 50% lower warehousing costs and reclaimed analyst hours, complete with a step-by-step blueprint.

    • 50% lowerdata-warehouse spend for the client, through elastic sizing and query optimisation (indicative)
    • 80% fastertime to insight, with average turnaround cut from 2 days to 3 hours (indicative)
    • < 60 secondsrebuilds for multi-billion-row fact tables in dbt pipelines (indicative)

    Stack

    • Snowflake
    • Amazon S3
    • Amazon EC2
    • GitLab CI/CD
    • Docker
    • dbt Core
    • Fivetran
    • Census
    • Power BI
    • Metabase

  3. 03

    End-to-end reverse ETL: Snowflake to Salesforce (lightweight, config-driven)

    Sector
    B2B SaaS
    Duration
    Delivered as a single focused build

    A lightweight, config-driven reverse ETL pipeline from Snowflake to Salesforce that saved about $10,000 a year in tooling fees, cut sync latency by roughly 90% via a direct bulk path, and gave engineering clear logs and auditability.

    • $10,000/yearsaved in managed reverse ETL tooling fees, for these syncs (indicative)
    • ~90% lowersync latency, via the direct Snowflake-to-Salesforce bulk path (indicative)
    • 1 YAML fileto add a new sync, with no application change and a reviewable Git diff (indicative)

    Stack

    • Python
    • Snowflake
    • Snowpark
    • Salesforce REST API
    • Salesforce Bulk API
    • YAML
    • Docker
    • GitLab CI
    • AWS Secrets Manager