The modern MLOps framework

Deployment e ciclo di vita in MLOps

Nemanja Radojkovic

Senior Machine Learning Engineer

About me

 

Profile picture

 

  • Senior Machine Learning Engineer
  • Experience developing, deploying, and maintaining Machine Learning models in production
  • Current focus: ML deployment and life-cycling
1 www.linkedin.com/in/radojkovic
Deployment e ciclo di vita in MLOps

Conceptual course

YES: High-level understanding

conceptual flowchart

NO: Hands-on skills

coding hands-on

Deployment e ciclo di vita in MLOps

MLOps

Machine Learning Operations

  • Principles
  • Practices
  • Tools

Goal: ML workflows and services

  • Automated
  • Reproducible
  • Integrated
Deployment e ciclo di vita in MLOps

ops part 1

Deployment e ciclo di vita in MLOps

ops part 2

Deployment e ciclo di vita in MLOps

mlops from dev

Deployment e ciclo di vita in MLOps

Chapter I

Value and necessity of MLOps

value

Model life cycle stages

life cycle

 

Components of the MLOps framework

building blocks

Deployment e ciclo di vita in MLOps

devops only

Deployment e ciclo di vita in MLOps

devops with mlops

Deployment e ciclo di vita in MLOps

High cost of ML without the Ops

 

USUAL STARTING POINT

  • Manual ML workflows
  • Manual deployment
  • Ad hoc monitoring

 

=> Accumulation of technical debt

Deployment e ciclo di vita in MLOps

Technical Debt

the implied cost of additional rework caused by choosing an easy (limited) solution now instead of using a better approach that would take longer

~ Wikipedia[1]

 

Famous Google paper on the topic:

"Machine Learning: The high-interest credit card of technical debt"[2]

 

MORE TIME AND MODELS DEPLOYED

  • More technical debt and model risk
  • Process slower, more frustrating and error prone
  • Ability to deliver value impaired
1 https://en.wikipedia.org/wiki/Technical_debt 2 https://research.google/pubs/pub43146/
Deployment e ciclo di vita in MLOps

ML workflows

  • Data collection and preparation
  • Data-labeling
  • Model selection
  • Model training
  • Model packaging
  • Model deployment
  • Model monitoring and maintenance

 

ML workflow automation == MLOps maturity

Deployment e ciclo di vita in MLOps

MLOps implemented

  • Automated
  • Fast
  • Reproducible
  • Explainable
Deployment e ciclo di vita in MLOps

we are here

Deployment e ciclo di vita in MLOps

non-linear

Deployment e ciclo di vita in MLOps

linear

Deployment e ciclo di vita in MLOps

spotlight 1

Deployment e ciclo di vita in MLOps

all eyez on me

Deployment e ciclo di vita in MLOps

Deployment e ciclo di vita in MLOps

Coming up!

 

Model life-cycling

life cycle

MLOps architecture

conceptual flowchart

Deployment e ciclo di vita in MLOps

Let's practice!

Deployment e ciclo di vita in MLOps

Preparing Video For Download...