Scribd
Upload
63 views23 pages

An introduction to AIMLOps

Introduction to AI MLOPS in Data Science

Uploaded by

santhoshnandigama
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
63 views23 pages

An introduction to AIMLOps

Introduction to AI MLOPS in Data Science

Uploaded by

santhoshnandigama
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 23

An

introduction to
AI/MLOps
SUNNY BHAVEEN CHANDRA | INEURON.AI
Contents
Part 1: Introduction Part 2: DevOps Vs Part 3: Steps in
AI/MLOps AI/MLOps

Objectives DevOps Generic Steps


Challenges Key Differences Workflow
Hidden Technical
Debt in ML Systems.
Part 1:

Introduction
01
Objectives 01
MLOps is an ML engineering culture and practice
that aims at unifying
ML system development (Dev) and
ML system operation (Ops)

02
Objectives 02
Techniques for implementing and automating -
continuous integration (CI),
continuous delivery (CD), and
continuous training (CT)
for machine learning (ML) systems.

02
Objectives 03
MLOps advocates for -
automation and
monitoring
at all stages of the ML system development
process, including -
integration,
testing,
releasing,
deployment, and
02 infrastructure management.
Challenges 01
Building an integrated ML system and continuously
operate it in production with a vast array of the
surrounding infrastructure.

02
Challenges 02
To automate the process from beginning to end
while managing -
different teams
using different technologies and
follow different routines.
And also make them -
auditable and
reproducible.
02
Challenges 03
Various dependencies like -
data dependency,
model complexity,
reproducibility,
testing,
monitoring etc

These changes and dependencies in addition to


code must be controlled and integrated into the
software delivery process.
02
Hidden
Technical
Debt in
Machine
Learning
Systems.
-Sculley et al. in 2015 04
Part 2:

DevOps Vs AI/MLOps
05
DevOps
In the development and operation of large-scale
software systems, DevOps is a widespread approach.
Shortening development cycles, boosting deployment
velocity, and ensuring reliable releases are all
advantages of this strategy.
It uses two concepts in software system development to
get these benefits:

1. Continuous Integration (CI)


2. Continuous Delivery (CD)

DESIGNED BY VECTORJUICE | FREEPIK

06
ML systems
differ from Continuous Integration (CI)

other It's no longer only about testing and validating code


and components; data, data schemas, and models

software must also be tested and validated.

systems
07
ML systems
differ from Continuous Delivery (CD)

other It's no longer about a single software package or


service, but about a system (an ML training pipeline)

software
that automatically deploys another service (model
prediction service).

systems
07
ML systems
differ from Continuous training (CT)

other It's a novel attribute specific to machine learning


systems that deals with automatically retraining and

software serving models.

systems
07
Part 3:

Steps in
08

AI/MLOps
Generic
steps in 1

3
Data extraction

Data preparation
2

4
Data analysis

Model training
AI/ML 5 Model evaluation 6 Model validation

systems 7 Model serving 8 Model monitoring

09
0 leveL

10 REFERENCE | GCP
PCG | ECNEREFER
1 leveL

11
PCG | ECNEREFER
2 leveL

12
Stages of
CI/CD

13 REFERENCE | GCP
1 GCP

MLOps: Continuous delivery


and automation pipelines in

References
machine learning

2 martinFowler.com

Continuous Delivery for


Machine Learning

14
Contact us if there are any questions. Website www.iNeuron.ai

Phone
+91 87885 03778 ,
Number +91 6260 726 925

Email
[email protected]
14 Address

You might also like