Case Studies | SaaS B2B

Owning the model

Rekognition to SageMaker Migration

About

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

Challenge

The challenge had three focus areas, each part of moving from a managed service to an owned model.

Replacing a managed service with an owned model

The first focus was migrating away from Amazon Rekognition without losing capability, standing up a custom object detection model that the client controls and can keep improving.

Training on the client's own data

A custom model needs good data. The second focus was preparing the client's dataset, updating the bounding box annotations to suit the chosen model, and fine-tuning a SageMaker object detection model on it.

Serving predictions in real time

Recognition has to happen live in the product. The final focus was deploying the model to a SageMaker endpoint and invoking it through a real-time inference pipeline that returns results to the user as images come in.

Solution

We structured the migration as a short sequence of steps, from data through to a live endpoint.

Custom

SageMaker object detection model on the client's data

Real-time

Inference via a live SageMaker endpoint

Owned

Recognition moved off a managed service to the client's control

By the numbers:

  • Custom - SageMaker object detection model on the client's data
  • Real-time - Inference via a live SageMaker endpoint
  • Owned - Recognition moved off a managed service to the client's control
Changes

The engagement delivered the client's move from a managed recognition service to a model it owns. A SageMaker object detection model was fine-tuned on the client's own dataset, deployed to a real-time endpoint, and made callable through a Lambda-based inference pipeline, meeting the migration's success criteria and proving AWS as the environment for the client's continued development.

  • Migration completed — Product recognition moved from Amazon Rekognition to a custom model on Amazon SageMaker.
  • Model fine-tuned on real data — The object detection model was trained and optimised on the client's own annotated dataset.
  • Real-time inference delivered — The model was deployed to a SageMaker endpoint and invoked through a Lambda inference pipeline for live results.
  • Handover — The client left the engagement owning its recognition model and pipeline, with control over accuracy and cost and a clear path to keep improving it.

With the model in the client's own hands, the natural next steps (projected) would be to expand the training data to cover more products, retrain to sharpen accuracy, and tune the endpoint for the throughput the product needs as usage grows.

AWS Stack

Amazon SageMaker

For training, hosting and serving the custom object detection model.

AWS Lambda

For the inference pipeline that invokes the model endpoint and returns results.

Amazon API Gateway

For the real-time interface that carries image requests and responses.

Amazon S3

For storing the training dataset, annotations and model artefacts.

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.

We use cookies

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

See our Privacy Policy for details.