Overview

Braylock is a US digital lending and payments company offering personal loans, a consumer card, and a payments app. Fast growth left its data spread across loan origination, servicing, payments, fraud and risk, collections, marketing, and support systems, all feeding an aging warehouse through hand-coded nightly batch ETL. The same records showed up in different shapes across systems, pipelines broke whenever a source schema changed, and risk and portfolio reporting ran a day or more behind. Onboarding a new source took weeks, and engineers spent most of their time patching jobs instead of building. Because the data carried regulated PII and fed regulatory reporting, fragile pipelines with no reliable lineage were also a compliance exposure. Bacancy applied data warehouse automation to rebuild the stack on Snowflake with metadata-driven, version-controlled ELT, automated testing, column-level lineage, orchestration, and CI/CD. Masking and role-based access protected regulated data, and business teams moved from day-old reports to governed, same-day dashboards.

Technologies Used

Snowflake
Fivetran
dbt
Coalesce
Apache Airflow
Git

Project Highlights

checkmark

Legacy hand-coded ETL replaced with metadata-driven, automated ELT generated from reusable patterns

checkmark

Data warehouse automation applied across loan, payment, fraud, and risk data on Snowflake

checkmark

Regulated financial data protected with masking, role-based access, and column-level lineage

checkmark

Same-day risk and portfolio reporting delivered through governed self-serve dashboards

The Challenges

1

Hand-coded ETL broke on every source schema change, and maintaining brittle jobs consumed most of the engineering week, leaving little time for new risk and analytics work.

2

Nightly batch meant risk, fraud, and portfolio reporting ran a day or more behind, so decisions relied on stale numbers and fraud signals were acted on late.

3

Adding a new source or metric took weeks. With no version control, testing, lineage, or documentation, critical logic lived in a few people's heads and was hard to audit.

4

The data included regulated PII that could not be exposed during processing, and regulatory reporting required auditable, reproducible lineage the legacy stack could not produce.

Solutions by Bacancy

1

Bacancy’s data warehouse services team rebuilt ingestion with managed connectors and moved transformation into the warehouse, so raw loan, payment, and risk data lands in Snowflake reliably and models run where the compute is.

2

Our Snowflake consulting services team set up metadata-driven, version-controlled ELT with dbt and Coalesce, so pipelines are generated from reusable patterns, tested on every change, and documented with column-level lineage instead of hand-coded one by one.

3

We replaced brittle cron jobs with orchestrated pipelines in Apache Airflow, added automated retries, alerting, and CI/CD, and cut over from the legacy system through a parallel run with automated validation, so reporting never dropped during the switch.

4

We rebuilt access on the analytics layer with dynamic masking and role-based rules so each user saw only what their role allowed and kept lineage and documentation audit-ready so risk and compliance decisions could be justified in review.

Core Features

checkmark

Metadata-driven, automated ELT generated on Snowflake

checkmark

Version-controlled, tested transformations with column-level lineage and auto-documentation

checkmark

Orchestrated self-healing pipelines with retries, alerting, and CI/CD

checkmark

Governed self-serve dashboards with masked PII and auditable lineage

No. of Resource

04

No. of Resource

Time Frame

August 2025 to February 2026

Time Frame

Project Snapshot

Data Warehouse Automation

Outcomes

75% shorter nightly batch window, from roughly 8 hours to under 2, so data was ready before the business day started

70% fewer pipeline failures after retiring hand-coded ETL for tested, pattern-generated models

80% faster onboarding of new data sources, with setup timelines reduced from several weeks down to a few days

60% of engineering time reclaimed from pipeline maintenance and redirected to risk and analytics work

90% of routine risk and portfolio reports delivered the same day, down from a turnaround of a full day or more

100% of regulated PII masked with role-based access, and full column-level lineage available for audit and regulatory reporting

Technical Stack

Cloud data warehouse Snowflake
Data ingestion Fivetran
Data warehouse automation & transformation dbt, Coalesce
Orchestration Apache Airflow
Data quality & observability dbt tests, Elementary
CI/CD & version control Git, GitHub Actions
BI / dashboards Power BI
Security & governance Snowflake RBACDynamic data maskingColumn-level lineage

Experience With Bacancy

2500+ Projects Experienced Innovation with Bacancy!

Get access to an experienced team of developers and engineers from Bacancy, handpicked to ace your goals. Kickstart within 48 hours, no-risk trial.

Book a 30 min call

14+

Years of Business Experience

1458+

Happy Customers

12+

Countries with Happy Customers

1050+

Agile Enabled Employees

How Can We Help?