Share via

How to manage fine-tuned GPT models and prompt flow versions in Azure AI Foundry?

satish sama 0 Reputation points
2025-10-30T22:40:18.7633333+00:00

I’m currently using Azure AI Foundry to build a few custom GPT models fine-tuned for different business scenarios (customer support, content generation, etc.). I’ve also integrated Prompt Flow to test and evaluate different prompt versions.

My question is — what’s the best practice for:

Managing multiple fine-tuned model versions in a single AI Foundry project?

Keeping track of Prompt Flow updates tied to each model version?

Ensuring that only specific team members can deploy or modify production-ready models?

Right now, it feels a bit tricky to maintain version alignment between fine-tuned models, prompt flows, and deployed endpoints.

Azure AI Bot Service
Azure AI Bot Service
An Azure service that provides an integrated environment for bot development.

3 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. Azar 31,610 Reputation points MVP Volunteer Moderator
    2025-10-30T22:41:48.23+00:00

    Hi there

    Keep separate projects for dev/test/prod.

    Register fine-tuned models with clear versioning in the Model Catalog.

    Tag Prompt Flows with the model version they use.

    Control deployment via RBAC so only approved users can publish.

    If this helps kindly accpt the answer

    0 comments
  2. Aryan Parashar 3,380 Reputation points Microsoft External Staff Moderator
    2025-10-31T07:17:18.7966667+00:00

    Hi Satish,

    Thankyou for posting this on Microsoft Q&A Portal.

    Multiple fine-tuned model versions could be managed in the following ways:

    For Prompt Flow updates tied to each model version, you can create a custom tool integrated with your prompt flow using the following documentation:
    https://microsoft.github.io/promptflow/how-to-guides/develop-a-tool/create-and-use-tool-package.html

    To ensure that only specific team members can deploy or modify production-ready models, enable managed identity for the AI Foundry resource and assign the contributor role to specific team members as shown below:
    User's image

    Feel free to accept this as an answer.
    Thankyou for reaching out to The Microsoft Q&A Portal.

    0 comments
  3. Deleted

    This answer has been deleted due to a violation of our Code of Conduct. The answer was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.

    Comments have been turned off. Learn more

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.