How Disco Transformed Their ML Pipeline and Cut Cloud Costs with Borealis Solutions

Jun 5

Industry: E-commerce / AI Checkout Solutions
Company Size: Startup
Region: United States
Website: www.disconetwork.com

Introduction

Disco is changing the way e-commerce businesses collaborate. As a Shopify partner, their AI-powered network surfaces intelligent, personalized promotions at checkout—creating a win-win across their merchant ecosystem.

But as they scaled, they hit some familiar startup roadblocks: rising AWS costs, slow backend performance, and inconsistent DevOps practices that slowed their ability to iterate and innovate.

That’s where Borealis Solutions came in.

The Challenge

Disco approached us with a clear need: modernize their ML and DevOps foundations to support rapid growth. Their biggest pain points included:

Disjointed DevOps practices and lack of operational maturity
Long backend startup times impacting developer velocity
Siloed access and weak visibility into infrastructure performance
Fragmented data workflows across AWS Glue, Redshift, and Snowflake
Escalating AWS bills with no cost control strategy

Why Disco Chose Borealis

Disco chose Borealis Solutions because we bring:

Deep expertise in DevOps transformation and AWS enablement
Experience optimizing machine learning infrastructure at early-stage startups
A hands-on approach that embeds within fast-moving engineering teams
A proven ability to improve performance while reducing costs

Our Approach

We worked alongside Disco’s machine learning and platform engineering teams to deliver a comprehensive modernization initiative:

1. ML & Data Infrastructure Support

Built ML pipelines for experimentation and deployment
Supported orchestration across AWS Glue, Redshift, and Snowflake
Streamlined cross-platform data operations to reduce friction

2. DevOps Transformation

Diagnosed and fixed slow backend startup performance
Set up CI/CD pipelines for consistent delivery workflows
Rolled out dashboards, metrics, and structured alerting to boost observability

3. Security & Access Control

Applied IAM best practices to enforce granular, team-specific access
Centralized access control to eliminate fragmentation across services

4. Cost Optimization & Platform Enablement

Conducted a full AWS cost analysis and applied right-sizing techniques
Enabled Disco’s internal teams to confidently manage and scale their infrastructure

The Results

In just weeks, Disco shifted from reactive operations to owning their platform:

✅ Faster ML experimentation and deployment cycles
✅ Reduced backend application startup time, accelerating development
✅ Unified DevOps culture with full observability and CI/CD workflows
✅ Secure, scalable AWS environment with structured access policies
✅ Lower AWS monthly spend without compromising performance

“The Borealis team helped us go from reactive DevOps to owning our platform. They made our ML pipeline efficient, helped us get control of costs, and ensured we could grow without compromising performance or visibility.”
— Disco Engineering Leadership

What’s Next for Disco

With foundational improvements in place, Disco is now positioned to scale confidently:

Expanding automation across the ML lifecycle
Optimizing data platform integrations to fuel deeper analytics
Scaling backend infrastructure to support rapid merchant acquisition

Brendon Palmer