Case Studies | SaaS B2B

An AWS-native platform for AI ad generation

AI Advertising Platform Modernisation

About

A SaaS B2B business working with Cloud Combinator on AWS. The client is anonymised at their request.

Challenge

The challenge had four focus areas, each about turning an externally dependent stack into a platform Nouvel could scale and stand behind.

Removing external generative dependencies

The existing third-party data layer and external generative video provider had to be replaced with self-hosted equivalents on AWS, giving Nouvel control over its data, its models and its roadmap.

Scaling concurrently and across regions

The platform needed to run in both eu-west-2 (London) and us-west-1 (N. California) with quotas and latency parity in each, supporting GDPR-driven separation of EU and US customer workloads under SLA-backed API integrations.

Cost per generation

Every ad generation carried a third-party bill. Bringing model inference and video generation onto self-hosted, GPU-backed AWS compute was designed to reduce per-generation cost materially, with a queue-based fallback for the agency-tier workflow.

Production-ready foundations

The architecture had to be secure and observable from day one, built as code and prepared for a Well-Architected Framework Review, a Foundational Technical Review and an eventual AWS Marketplace listing.

Solution

We delivered the modernisation as a sequence of milestones, each with a clear deliverable and acceptance point, so Nouvel could see the platform take shape and validate it region by region.

2 regions

Active across London (eu-west-2) and N. California (us-west-1)

90,000+

Generations per month the platform is sized for (projected)

100%

Deployed as code via CloudFormation and SAM

By the numbers:

  • 2 regions - Active across London (eu-west-2) and N. California (us-west-1)
  • 90,000+ - Generations per month the platform is sized for (projected)
  • 100% - Deployed as code via CloudFormation and SAM
Changes

The modernised platform was delivered and demonstrated end to end across both deployment regions under representative concurrent load. Nouvel's external generative dependencies are replaced with self-hosted AWS equivalents, and the whole stack deploys from version-controlled infrastructure as code.

  • Self-hosted generationThe third-party data layer and external video provider are replaced by Aurora PostgreSQL, SageMaker-hosted Llama models and self-hosted video generation on ECS GPU compute.
  • Multi-region by designSymmetric deployments in eu-west-2 and us-west-1 support GDPR-driven separation of EU and US workloads, validated for multi-region failover.
  • Asynchronous and scalableStep Functions and SQS replace synchronous polling with a horizontally scalable, multi-tenant pipeline suited to concurrent load.
  • Lower cost per generationBringing inference and video onto self-hosted, GPU-backed AWS compute is designed to reduce per-generation cost materially versus the previous external provider.
  • Secure and observableLeast-privilege IAM, AWS Secrets Manager, AWS WAF, CloudWatch and X-Ray give the platform production-grade security and end-to-end visibility.
  • Self-improving modelsAn EventBridge-driven curation and SageMaker retraining loop refreshes the niche-specific model and promotes it via blue/green deployment.
  • HandoverNouvel received working sessions, walkthrough recordings and an architecture handover pack covering the IaC repositories, runbooks and next steps toward a Well-Architected and Foundational Technical Review.

With an AWS-native, multi-region foundation in place, Nouvel owns its data, its models and its generation pipeline end to end, and is positioned for SLA-backed enterprise contracts, a Well-Architected Framework Review and an eventual AWS Marketplace listing as the business grows.

AWS Stack

Amazon Bedrock

For concept and script generation with cross-region inference profiles.

Amazon SageMaker

For fine-tuning, hosting and retraining the niche-specific Llama models.

Amazon ECS

On EC2 GPU instances for self-hosted video generation.

AWS Step Functions

With Amazon SQS for asynchronous, end-to-end pipeline orchestration.

Amazon Aurora PostgreSQL

And Amazon S3 for the application database and media storage.

AWS Fargate

For the non-GPU scrape, audio-mux and stitching pipeline stages.

Amazon CloudWatch

And AWS X-Ray for logging, metrics, alarms and distributed tracing.

AWS CloudFormation

And SAM for provisioning the full stack across both regions as code.

YOU MIGHT LIKE

Related success stories

View all case studies

Case Studies | Insights

Utilising Language Recognition, Speed, and Enhanced Security to Make Social Media a Force for Good

  • Here, we take a detailed look at how the Cloud Combinator team collaborated with another cutting-edge AI service provider that provides intelligent systems to “make social media more social” for brands and users alike.
  • Arwen AI is a UK-based startup specialising in AI solutions to manage and enhance brands’ social media interactions. Founded in 2020 by Matt McGrory, Dr. David Cole, and Joel Bailey, Arwen. AI focuses on using AI to automatically detect and remove spam, toxic comments, and other unwanted content from social media platforms.
  • The team at Arwen have three core products. ‘Moderate’ is focused on identifying and removing toxic content from social media channels. ‘Engage’ helps brands identify and engage with meaningful conversations on social media, and ‘Customize’ allows brands to apply bespoke algorithms to their channels - creating an even more effective moderation and engagement.
Read more
CONTACT US

Ready to turn AI into impact?

We'll help you spot the highest-value opportunities, reduce risk around your first AI initiative, and define a clear path to results from day one.

Why talk to us:

Outcome-driven recommendations

AWS-recognised delivery expertise

Risk-aware AI adoption

Clear next step, not a sales pitch

Start with a focused 20-minute conversation about your goals — no pressure, no commitment.

This website uses cookies to enhance user experience and to analyze performance and traffic on our website.

See our Privacy Policy for details.