Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Taking ML to production - a journey
Search
Arnon Rotem-Gal-Oz
PRO
July 06, 2021
Technology
0
88
Taking ML to production - a journey
Go over some of the complexities of turning a machine learning solution to one used in production
Arnon Rotem-Gal-Oz
PRO
July 06, 2021
Tweet
Share
More Decks by Arnon Rotem-Gal-Oz
See All by Arnon Rotem-Gal-Oz
Brownfield Architecture transformations
arnonrgo
PRO
0
39
Software architecture 101
arnonrgo
PRO
0
640
Apache Spark - Overview
arnonrgo
PRO
0
31
Topics in Distributed Systems
arnonrgo
PRO
0
16
Docker & Kubernetes
arnonrgo
PRO
0
9
Data Security @ the personal level
arnonrgo
PRO
0
12
Microservices it's deja vu all over again
arnonrgo
PRO
0
10
Big Data in the Cloud - Welcome to cost oriented design
arnonrgo
PRO
0
8
Big Data Overview
arnonrgo
PRO
0
7
Other Decks in Technology
See All in Technology
データベース08: 実体関連モデルとは?
trycycle
0
100
AthenaとStep Functionsで簡単ETLオーケストレーション #midosuji_tech
nayuts
0
330
「メタスキル」を意図的に使おう ~自律的なチームを育むマネージャーのセルフマネジメント~/Use Metaskills consciously
ohnoeight
1
520
FFMでJITコンパイラを作ってみた
yasuenag
0
130
C / C++ - language
pohjus
1
260
Optimising Scala 3 Build Times - Scala Matsuri
bishabosha
0
120
オンプレADを昇降格したらAWS上のシステムのいろんなところが通信できなくなった話
hamijay_cloud
0
280
JVM言語でもできる、競技プログラミング
dhirabayashi
0
140
社内LLMハッカソン2024発表資料
streamwest1629
0
110
Goでテストをしやすくするためにやったこと
kazukihayase
1
530
成長し続けるTVerサービスを支える オブザーバビリティとカスタマーサポート
techtver
PRO
2
700
不要な DNS リソースレコードは消そう / Delete unused DNS records
wa6sn
4
3.1k
Featured
See All Featured
Responsive Adventures: Dirty Tricks From The Dark Corners of Front-End
smashingmag
247
20k
Speed Design
sergeychernyshev
2
77
In The Pink: A Labor of Love
frogandcode
138
21k
Gamification - CAS2011
davidbonilla
77
4.7k
CSS Pre-Processors: Stylus, Less & Sass
bermonpainter
353
28k
Automating Front-end Workflow
addyosmani
1359
200k
What’s in a name? Adding method to the madness
productmarketing
PRO
18
2.8k
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
34
6.2k
Raft: Consensus for Rubyists
vanstee
133
6.4k
The Cost Of JavaScript in 2023
addyosmani
25
4.1k
ParisWeb 2013: Learning to Love: Crash Course in Emotional UX Design
dotmariusz
104
6.7k
A designer walks into a library…
pauljervisheath
201
23k
Transcript
Taking ML to production - A journey Arnon Rotem-Gal-Oz
Mental model of the probelm Admission Intubation Alert >6 hours
Challlenge 1 defining the problem
A Perfect Representation of the Machine Learning Cycle from start
to end | Image Source: MLOps (Published under Creative Commons Attribution 4.0 International Public License and used with attribution (“INNOQ”))
None
Challenge 2 – how we measure
Challenge 3 Data quality
None
Challenge 5 different types of data Model(s) text time series
categorical
Challenge 6 labeling Admission Intubation
Challenge 7 dealing with imballance
Challenge 8 Model experimentation cycle
Modeling
Challenge 9 – Overfit
None
Moving to production…
Challenge 10 – model degredation in production Theory Reality
Challenge11 – Is it really generalized?
Challenge 12 model validation and verification
The road to production…