Regarding the ML Engineer we are looking for at Thoucentric, these are the activities we need them to do:
• Improve the draft model pipelines between Sagemaker Model Registry and Sagemaker Endpoints, for deployment on QA, PROD and deletion of rejected models on QA.
• Implement model lineage as model cards in Sagemaker Model Registry, detailing which data version and Bitbucket tag was used to train the registered model.
• Implement multi-model Sagemaker Model Endpoints and integrate those with Azure AD (SIGA).
• Implement Model Monitoring of models deployed on QA and PROD. (We are currently exploring Sagemaker Model Monitoring, but it has problems monitoring multi-model endpoints, so the decision is yet to be finalized.)
• Implement bespoke model monitoring required by Legal following gap analysis against Sagemaker Model Monitor solution. (This activity will be done later this year.)
• Activate model monitoring for models already running in PROD. (This activity will be done later this year.)
• Co-create with architects & DevOps how model deployment, model monitoring etc will work following the DEEP migration in 2025.
Any Graduate