← Work

2023 · AI infrastructure

A serverless fine-tuning pipeline for personalized AI images

a consumer AI photo app startup · Ai infrastructure engineer

The problem

A photo app wanted users to upload their own photos and get new AI images of themselves. That needs a model trained on each person, not a single call to a hosted service.

What I built

I built a serverless pipeline that takes the uploaded photos, trains a personal model on cloud machines and stores the result, controlled by a workflow.

The result

The app could offer personal AI images without building a training system itself.

Main tools

  • AWS Lambda
  • API Gateway
  • AWS Step Functions
  • Amazon SageMaker
  • Amazon S3

This is the short version. The full case study is available on request.

Ask for the full case study →