From 888e77b3b9588b5c94b3766165d093429ff78c89 Mon Sep 17 00:00:00 2001 From: Hema Ramamoorthy <93590728+hema-dc@users.noreply.github.com> Date: Tue, 6 Aug 2024 19:52:58 +0100 Subject: [PATCH] Create START.md --- START.md | 81 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 81 insertions(+) create mode 100644 START.md diff --git a/START.md b/START.md new file mode 100644 index 0000000..6a8c5c4 --- /dev/null +++ b/START.md @@ -0,0 +1,81 @@ +# Effective ML Deployment + +ML Deployment - the process of deploying an AI/ML model from PoC to Production + +# Key Offerings +image + +## Startups +1. Gather / Assess + + Gather + * The requirements, goals and objectives + * Understand the AI/ML use case(s) + * Understand the current infrastructure requirements + * Understand the deployment expectations (Budget, timescales, resources) + + Assess + * The current AI/ML model + * The maturity of the code + * The current available infrastructure + * The current data pipeline (if any) + * The Data dependencies + * The size of the current dataset + * If current data is real or synthetic + * The data source for model in production + * Skills availability within the organisation +3. Plan / Design + * Design the potential deployment pipeline + * Plan the deployment strategy + * Plan the metrics to be monitored +4. Implement + * Deploy the model + * Review the deployment strategy +5. Evaluate / Review + * Monitor the metrics of the model + * Observability - Model health, Data health and Service health + +## Large Organisations +1. Gather / Assess + + Gather + * The requirements, goals and objectives + * Understand the AI/ML use case(s) and their connectivity to organisational strategy + * Understand the current infrastructure requirements + * Understand the deployment expectations (Budget, timescales, resources) + + Assess + * The current AI/ML model + * Past AI/ML model (if any) already in production + * The maturity of the code + * The MLOps Maturity level + * The Scalability requirements + * The current available infrastructure + * The current data pipeline (if any) + * The Data dependencies + * The size of the current dataset + * If current data is real or synthetic + * The data source for model in production + * Skills availability within the organisation +3. Plan / Design + * Design the potential deployment pipeline + * Plan the deployment strategy + * Plan the metrics to be monitored +4. Execute + * Deploy the model + * Review the deployment strategy +5. Evaluate / Review + * Monitor the metrics of the model + * Observability - Model health, Data health and Service health + +# Topics Covered +1. Gathering Requirements +2. Assessing the organisation's ML Readiness - [Maturity Assessment](https://github.com/hema-dc/ML-Deployment/blob/main/Offerings/Maturity%20Assessment.md) +3. [Best Practices](https://github.com/hema-dc/ML-Deployment/blob/main/Offerings/Best%20Practices.md) +4. Horizon Scan of available MLOps Tools +5. ML Deployment +6. [Deployment Strategies](https://github.com/hema-dc/ML-Deployment/blob/main/Offerings/Strategies.md) +7. [MLOps](https://github.com/hema-dc/ML-Deployment/blob/main/Offerings/MLOps.md) +8. MLOps Tools +9. [ML Monitoring and Observability](https://github.com/hema-dc/ML-Deployment/blob/main/Offerings/Observability.md) +10. [MLOps Skills](https://github.com/hema-dc/ML-Deployment/blob/main/Offerings/MLOps%20Skills.md)