SlideShare a Scribd company logo
1 of 35
Download to read offline
MLOps Journey at Swisscom
AIUseCases,ArchitectureandFutureVision
Joana Soares Machado
Maxime Darçot
24.02.2023
Outline
2
• Overview: AI Use Cases at Swisscom
• Challenges
• AI Use Cases Deep Dive
• MLOps Principles
• Future Outlook: The AWS Move
MLOps
journey
at
Swisscom
Overview
AI Use Cases at Swisscom
3
Infrastructure Analytics
Anomaly Detection, Network Optimization,
Customer Centric Monitoring, Mobility Insights
Conversational AI
Conversational AI for Swisscom’s Products &
Services, Innovation & Research
Data Services
Data Lake & Warehouse, Enablers, Data and
Analytics Tools in Self-service
B2B Analytics
Business Insights about B2B customers and
Process Performance, KPIs definition
Mass Market Analytics
SC Touchpoints Insights, Pricing Simulations for
Offers, Recommender and Offer System to Sales
Data, Analytics
& AI Business Analytics
Analytics solutions for Sales, Logistics, Finance,
Controlling and HCM, Management Reporting
4
MLOps
journey
at
Swisscom
5
350
experts in
Data, Analytics & AI
Skill Overview
• Data Science / AI Solution Design
• Software Development
• Data Engineering
• Natural Language Processing
• Business Engineering & Design
• Reporting & Analysis Specialists
• User Experience Design
Data & Infrastructure Overview
• >660 servers' on-premise infrastructure
• >40 Data, Analytics and AI services &
platforms & > 150 tools
• 10.73 Peta Bytes (PB) data
• 8.17 Mio. real-time messages per second
• 4’756 monthly active users of dashboard
tools
MLOps
journey
at
Swisscom
Challenges
6
7
MLOps
journey
at
Swisscom
Data Regulations
CH-only, on-prem
Data Formats
Tabular, time series
Challenges: Diversity of...
Scalability Req.
Nb of models, SLAs
• Millions of models @INI
(due to anomaly detection)
• 100s of models (@B2C & @B2B)
• Customer facing applications
Tech Stacks
Clouds, legacy
• k8s on AWS & SBD
• Internal Application
Cloud (CF)
AI Applications
NLP, ASR, Real-time
• 40 product instances @DNA
• 350 Users @DNA
• ML Engineers
• Business Analysts
• 3000 Users @Swisscom
Deep Dive
AI Use Cases
8
9
Digital Marketplace
35M API calls/day
Software Applications
300M Function calls/day
Business Processes
80K Processes/day
Network Infrastructure
20B Network interactions/day
MLOps
journey
at
Swisscom
Anomaly Detection Use Cases at Swisscom
Source: Towards Data Science
Time series Anomaly Detection
10
Data Sourcing
(Time series)
Real-time
Anomaly
Detection
UI + Alerting
Outage
Detection Insights
Feedback
MLOps
journey
at
Swisscom
Anomaly Detection Pipeline
/
Daily
Data Preparation Training + Batch
Prediction
11
MLOps
journey
at
Swisscom
Anomaly Detection Pipeline: Open Source Tools
CI/CD Version Control System Workflow Orchestration
ML Resource
Infrastructure
ML Metadata Store ML Model Serving Pipeline Monitoring
• Measure lags and pipeline availability.
• Different aspects of quality:
− Training data validation;
− Deploying models with good accuracy;
− Logging information about models and
errors, for transparency and reproducibility.
• SLI/SLO alerting: target over time period.
12
MLOps
journey
at
Swisscom
MLOps: Managing the ML Lifecycle
Source: Sensu
13
MLOps
journey
at
Swisscom
Experiment Tracking with Mlflow
14
MLOps
journey
at
Swisscom
Monitoring over time – Model Accuracy
15
MLOps
journey
at
Swisscom
Monitoring over time – Requests and Processing Time
16
MLOps
journey
at
Swisscom
Conversational AI – Dialog flow
16
Language: EN
Intent: Customer has a technical issue
Topics: Internet
Hello
I see that your internet connection
has been blocked because you
have 3 outstanding bills
Do you wish to get more time to
pay these bills?
Hello and welcome to Swisscom,
how may we help you today?
My internet stopped working
Yes please
Would you be able to pay by the
end of next month?
Sure, even by the 15th
Alright I’ve unblocked your
connection, you have until April
15th to pay these 3 bills.
Thank you
Intent: Customer confirms
Topics: None
Intent: Customer confirms
Topics: None
Intent: Customer thanks us
Topics: None
Language
Detection
Intent & Topic
classifiers
General Dialog
Manager
Billing APIs
Intent: Customer greets us
Topics: None
Background call to billing API -> Customer is blocked
Entity
Recognition
Entity: 15th (Date)
Domain
classifier
Domain: None
Domain: Technical support
Domain: Billing
Technical
Support Dialog
Manager
Billing Dialog
Manager
General Dialog
Manager
Background call to billing API -> Change due dates
17
Speech
Recognition
Language
Detection
Entity
Recognition
Machine
Translation
Intent & Topic
classifiers
Retrieval
System
Domain
classifier
…
Monitoring Reporting
Model Storage
CI / CD Error alerting
Secure data
storage
Flow design
tool
Data labeling
tools
End to end
testing tool
Training
pipelines
Billing APIs
CRM
…
Hotline
Whatsapp
TV Box
E-mail
Letter / Fax
...
Input APIs
Dialog
Managers
Conversational AI stack
18
Input
Channels
AI/ML layer
Runtime foundations layer
Offline tools
layer
Input APIs Integration
layer
Dialog
Managers
Conversational AI stack
19
Speech
Recognition
Language
Detection
Entity
Recognition
Machine
Translation
Intent & Topic
classifiers
Retrieval
System
Domain
classifier
…
Monitoring Reporting
Model Storage
CI / CD Error alerting
Secure data
storage
Flow design
tool
Data labeling
tools
End to end
testing tool
Training
pipelines
Billing APIs
CRM
…
Hotline
Whatsapp
TV Box
E-mail
Letter / Fax
...
Input APIs
Dialog
Managers
Conversational AI – ML Operations
20
Intent & Topic
classifiers
Domain
classifier
Model Storage
CI / CD
Secure data
storage
Data labeling
tools
Training
pipelines
Conversational AI – ML Operations
Intent & Topic
classifiers
Domain
classifier
Model Storage
CI / CD
Secure data
storage
Data labeling
tools
Training
pipelines
21
Intent & Topic
classifiers
Domain
classifier
Model Storage
CI / CD
Secure data
storage
Data labelling
tools
Training
pipelines
Conversational AI – ML Operations
22
MLOps
journey
at
Swisscom
Conversational AI: MLOps Tools
CI/CD Version Control System Workflow Orchestration
ML Resource
Infrastructure
Secure Data Storage AI Labelling Tool ML Model Serving
MLOps Principles
23
24
(ML) Product Lifecycle
SW
Product
1.
Development
2.
Integration
3.
Testing
4.
Deployment
5.
Feedback
6.
Monitoring
7.
Operations
ML Product
2.
Data
Preparation
1.
Data
Collection
3.
Data
Integration
4.
Data
Transformation
5.
Model Training
6.
Model
Registration
7.
Model
Deployment
8.
Model
Evaluation
9.
Model
Monitoring
A. Data
Engineering
B. Model
Engineering
C. Model
Operations
MLOps
journey
at
Swisscom
25
Tools for the MLOps Principles over the Data Product Lifecycle
Principle
Collaboration [P5]
Reproducibility [P3]
CI/CD automation [P1]
Workflow orchestration [P2]
Continuous ML training/evaluation [P6]
Feedback loops [P9]
ML metadata tracking/logging [P7]
Continuous monitoring [P8]
Versioning of data, code, model [P4]
Adapted from: Machine Learning Operations (MLOps): Overview, Definition, and Architecture
C. Model
Operations
A. Data
Engineering
B. Model
Engineering
Version
Control
System
P4, P5
CI/CD
Component
P1, P6, P9
ML Resource
Infrastructure
P6
Feature Store
System
P3, P4
Workflow
Orchestration
Component
P2, P3, P6
ML Model
Registry
P3, P4
Monitoring
Component
P8, P9
ML Metadata
Store
P4, P7
ML Model
Serving
Component
P1
MLOps
journey
at
Swisscom
MLOps
journey
at
Swisscom
26
MLOps Harmonization – A Central Solution at Swisscom?
Standardized solution
Easy to start new project
No need to reinvent the wheel every time
Collaborate
and exchange best practices
• How much of the requirements does it cover?
• How good is the user experience?
• What is outlook on future requirements?
• What is outlook on future UX ?
27
MLOps Solutions - Evaluation Dimensions
• Development & Integration Costs.
• Operations & Maintenance Costs.
• Infrastructure (incl. Licenses) Costs.
• Outlook on cost evolution.
User Needs & UX Costs & Efficiency
MLOps
journey
at
Swisscom
Personas / Roles
• Yes-Code vs No-Code:
− Data Engineer,
Data Scientist,
(ML)Ops Engineer,
SW Engineer.
− Business Analyst,
Product Manager.
Future Outlook
The AWS Move
28
29
We want to converge our warehouses & Big Data lake into a hybrid lake house
architecture
From
on prem only
Data Lake & Warehouse
Warehouses
Swisscom Data
Warehouse
ONDS
B2B DWH
Swisscom
Big Data Platform
Data Lake
Hybrid Data Factory
+
To
a hybrid
Data Lakehouse
One
Data Platform
MLOps
journey
at
Swisscom
30
One Data Platform – High-level Architecture
Store &
Process
Applications
& Data
Functions
Exposure
Gatekeeper
Gatekeeper
MLOps DWH
Analytic
s
e.g. Time
Series
Custom
Apps
e.g.
Advanced
Processing
Ingestion
Storage
Transformations
Ingestion
Storage
Transformations
Data Hub
On Premise (for very specific workloads) AWS (whenever possible)
Gatekeeper
Gatekeeper
Data Access
DWH
Analytics
e.g. Time
Series
MLOps
Custom
Apps
e.g.
Advanced
Processing
real-time
/
batch
real-time
/
batch
CI
/CD
Data
Management
Foundation
3 2 1
4
6
5
7 8
9
4
MLOps
journey
at
Swisscom
• AWS Sagemaker
• All-in-one, yet modular
• Including no-code option
• Amazon MWAA
• Managed Airflow (orchestration)
• DataRobot
• All-in-one
• Including no-code option
31
MLOps - Commercial Product Candidates
• All-in-one
• Databricks, Snowflake
• Single components
• Fiddler AI (model performance management)
• Arize (observability)
• CometML (model management & monitoring)
• Astronomer (orchestration)
Concrete Options Other Possibilities
MLOps
journey
at
Swisscom
32
MLOps – Open Source Tools
MLOps
journey
at
Swisscom
Pros
• Free solutions, community built.
• Continuously improved and well tested.
• Independent solutions for each component:
• Fully customizable and modular.
• Easy switch when better options available.
• Infrastructure agnostic (Hybrid cloud).
Cons
• Needs to be managed by Swisscom engineers.
• May be more costly than “managed”solutions in
terms of infrastructure.
• Most solutions don’t offer a “no-code”option.
33
MLOps – Hybrid Solution (AWS & Open source)
MLOps
journey
at
Swisscom
Bringing together the best of both worlds?
• Huge diversity of AI use cases.
• Need for harmonization in the MLOps area.
• The AWS move brings benefits but also questions.
• A hybrid setup will likely be a main part of the solution.
• This is work in progress, our MLOps journey is not at an
end yet.
Summary
34
MLOps
journey
at
Swisscom
35
Thank you!
Maxime Darçot
maxime-darcot
Joana Soares Machado
joana-soares-machado

More Related Content

What's hot

MLOps Bridging the gap between Data Scientists and Ops.
MLOps Bridging the gap between Data Scientists and Ops.MLOps Bridging the gap between Data Scientists and Ops.
MLOps Bridging the gap between Data Scientists and Ops.Knoldus Inc.
 
MLOps - The Assembly Line of ML
MLOps - The Assembly Line of MLMLOps - The Assembly Line of ML
MLOps - The Assembly Line of MLJordan Birdsell
 
Pythonsevilla2019 - Introduction to MLFlow
Pythonsevilla2019 - Introduction to MLFlowPythonsevilla2019 - Introduction to MLFlow
Pythonsevilla2019 - Introduction to MLFlowFernando Ortega Gallego
 
What’s New with Databricks Machine Learning
What’s New with Databricks Machine LearningWhat’s New with Databricks Machine Learning
What’s New with Databricks Machine LearningDatabricks
 
Klassifizierung von Versicherungsschäden – AI und MLOps bei der Mobiliar
Klassifizierung von Versicherungsschäden – AI und MLOps bei der MobiliarKlassifizierung von Versicherungsschäden – AI und MLOps bei der Mobiliar
Klassifizierung von Versicherungsschäden – AI und MLOps bei der MobiliarBATbern
 
Productionzing ML Model Using MLflow Model Serving
Productionzing ML Model Using MLflow Model ServingProductionzing ML Model Using MLflow Model Serving
Productionzing ML Model Using MLflow Model ServingDatabricks
 
Introdution to Dataops and AIOps (or MLOps)
Introdution to Dataops and AIOps (or MLOps)Introdution to Dataops and AIOps (or MLOps)
Introdution to Dataops and AIOps (or MLOps)Adrien Blind
 
Vertex AI - Unified ML Platform for the entire AI workflow on Google Cloud
Vertex AI - Unified ML Platform for the entire AI workflow on Google CloudVertex AI - Unified ML Platform for the entire AI workflow on Google Cloud
Vertex AI - Unified ML Platform for the entire AI workflow on Google CloudMárton Kodok
 
AWS Monitoring & Logging
AWS Monitoring & LoggingAWS Monitoring & Logging
AWS Monitoring & LoggingJason Poley
 
How to Utilize MLflow and Kubernetes to Build an Enterprise ML Platform
How to Utilize MLflow and Kubernetes to Build an Enterprise ML PlatformHow to Utilize MLflow and Kubernetes to Build an Enterprise ML Platform
How to Utilize MLflow and Kubernetes to Build an Enterprise ML PlatformDatabricks
 
Data & Analytics ReInvent Recap [AWS Basel Meetup - Jan 2023].pdf
Data & Analytics ReInvent Recap [AWS Basel Meetup - Jan 2023].pdfData & Analytics ReInvent Recap [AWS Basel Meetup - Jan 2023].pdf
Data & Analytics ReInvent Recap [AWS Basel Meetup - Jan 2023].pdfChris Bingham
 
Introducing Databricks Delta
Introducing Databricks DeltaIntroducing Databricks Delta
Introducing Databricks DeltaDatabricks
 
Using MLOps to Bring ML to Production/The Promise of MLOps
Using MLOps to Bring ML to Production/The Promise of MLOpsUsing MLOps to Bring ML to Production/The Promise of MLOps
Using MLOps to Bring ML to Production/The Promise of MLOpsWeaveworks
 
Productionizing Machine Learning with a Microservices Architecture
Productionizing Machine Learning with a Microservices ArchitectureProductionizing Machine Learning with a Microservices Architecture
Productionizing Machine Learning with a Microservices ArchitectureDatabricks
 

What's hot (20)

MLOps Bridging the gap between Data Scientists and Ops.
MLOps Bridging the gap between Data Scientists and Ops.MLOps Bridging the gap between Data Scientists and Ops.
MLOps Bridging the gap between Data Scientists and Ops.
 
MLOps - The Assembly Line of ML
MLOps - The Assembly Line of MLMLOps - The Assembly Line of ML
MLOps - The Assembly Line of ML
 
Pythonsevilla2019 - Introduction to MLFlow
Pythonsevilla2019 - Introduction to MLFlowPythonsevilla2019 - Introduction to MLFlow
Pythonsevilla2019 - Introduction to MLFlow
 
Building your Datalake on AWS
Building your Datalake on AWSBuilding your Datalake on AWS
Building your Datalake on AWS
 
What’s New with Databricks Machine Learning
What’s New with Databricks Machine LearningWhat’s New with Databricks Machine Learning
What’s New with Databricks Machine Learning
 
MLOps for production-level machine learning
MLOps for production-level machine learningMLOps for production-level machine learning
MLOps for production-level machine learning
 
Klassifizierung von Versicherungsschäden – AI und MLOps bei der Mobiliar
Klassifizierung von Versicherungsschäden – AI und MLOps bei der MobiliarKlassifizierung von Versicherungsschäden – AI und MLOps bei der Mobiliar
Klassifizierung von Versicherungsschäden – AI und MLOps bei der Mobiliar
 
Productionzing ML Model Using MLflow Model Serving
Productionzing ML Model Using MLflow Model ServingProductionzing ML Model Using MLflow Model Serving
Productionzing ML Model Using MLflow Model Serving
 
Machine Learning Operations & Azure
Machine Learning Operations & AzureMachine Learning Operations & Azure
Machine Learning Operations & Azure
 
Introdution to Dataops and AIOps (or MLOps)
Introdution to Dataops and AIOps (or MLOps)Introdution to Dataops and AIOps (or MLOps)
Introdution to Dataops and AIOps (or MLOps)
 
Vertex AI - Unified ML Platform for the entire AI workflow on Google Cloud
Vertex AI - Unified ML Platform for the entire AI workflow on Google CloudVertex AI - Unified ML Platform for the entire AI workflow on Google Cloud
Vertex AI - Unified ML Platform for the entire AI workflow on Google Cloud
 
AWS Monitoring & Logging
AWS Monitoring & LoggingAWS Monitoring & Logging
AWS Monitoring & Logging
 
How to Utilize MLflow and Kubernetes to Build an Enterprise ML Platform
How to Utilize MLflow and Kubernetes to Build an Enterprise ML PlatformHow to Utilize MLflow and Kubernetes to Build an Enterprise ML Platform
How to Utilize MLflow and Kubernetes to Build an Enterprise ML Platform
 
Data & Analytics ReInvent Recap [AWS Basel Meetup - Jan 2023].pdf
Data & Analytics ReInvent Recap [AWS Basel Meetup - Jan 2023].pdfData & Analytics ReInvent Recap [AWS Basel Meetup - Jan 2023].pdf
Data & Analytics ReInvent Recap [AWS Basel Meetup - Jan 2023].pdf
 
Introducing Databricks Delta
Introducing Databricks DeltaIntroducing Databricks Delta
Introducing Databricks Delta
 
From Monolith to Microservices
From Monolith to MicroservicesFrom Monolith to Microservices
From Monolith to Microservices
 
Why to Cloud Native
Why to Cloud NativeWhy to Cloud Native
Why to Cloud Native
 
Using MLOps to Bring ML to Production/The Promise of MLOps
Using MLOps to Bring ML to Production/The Promise of MLOpsUsing MLOps to Bring ML to Production/The Promise of MLOps
Using MLOps to Bring ML to Production/The Promise of MLOps
 
MLOps with Kubeflow
MLOps with Kubeflow MLOps with Kubeflow
MLOps with Kubeflow
 
Productionizing Machine Learning with a Microservices Architecture
Productionizing Machine Learning with a Microservices ArchitectureProductionizing Machine Learning with a Microservices Architecture
Productionizing Machine Learning with a Microservices Architecture
 

Similar to MLOps journey at Swisscom: AI Use Cases, Architecture and Future Vision

5 Years Of Building SaaS On AWS
5 Years Of Building SaaS On AWS5 Years Of Building SaaS On AWS
5 Years Of Building SaaS On AWSChristian Beedgen
 
API and Big Data Solution Patterns
API and Big Data Solution Patterns API and Big Data Solution Patterns
API and Big Data Solution Patterns WSO2
 
Achieving Massive Concurrency & Sub-second Query Latency on Cloud Warehouses ...
Achieving Massive Concurrency & Sub-second Query Latency on Cloud Warehouses ...Achieving Massive Concurrency & Sub-second Query Latency on Cloud Warehouses ...
Achieving Massive Concurrency & Sub-second Query Latency on Cloud Warehouses ...Alluxio, Inc.
 
DataLive conference in Geneva 2018 - Bringing AI to the Data
DataLive conference in Geneva 2018 - Bringing AI to the DataDataLive conference in Geneva 2018 - Bringing AI to the Data
DataLive conference in Geneva 2018 - Bringing AI to the DataSasha Lazarevic
 
Big Data Analytics Platforms by KTH and RISE SICS
Big Data Analytics Platforms by KTH and RISE SICSBig Data Analytics Platforms by KTH and RISE SICS
Big Data Analytics Platforms by KTH and RISE SICSBig Data Value Association
 
Blueprint Series: Architecture Patterns for Implementing Serverless Microserv...
Blueprint Series: Architecture Patterns for Implementing Serverless Microserv...Blueprint Series: Architecture Patterns for Implementing Serverless Microserv...
Blueprint Series: Architecture Patterns for Implementing Serverless Microserv...Matt Stubbs
 
MindSphere: The cloud-based, open IoT operating system. Damiano Manocchia
MindSphere: The cloud-based, open IoT operating system. Damiano ManocchiaMindSphere: The cloud-based, open IoT operating system. Damiano Manocchia
MindSphere: The cloud-based, open IoT operating system. Damiano ManocchiaData Driven Innovation
 
Von der Zustandsüberwachung zur vorausschauenden Wartung
Von der Zustandsüberwachung zur vorausschauenden WartungVon der Zustandsüberwachung zur vorausschauenden Wartung
Von der Zustandsüberwachung zur vorausschauenden WartungPeter Schleinitz
 
Feature Store as a Data Foundation for Machine Learning
Feature Store as a Data Foundation for Machine LearningFeature Store as a Data Foundation for Machine Learning
Feature Store as a Data Foundation for Machine LearningProvectus
 
Platform governance, gestire un ecosistema di microservizi a livello enterprise
Platform governance, gestire un ecosistema di microservizi a livello enterprisePlatform governance, gestire un ecosistema di microservizi a livello enterprise
Platform governance, gestire un ecosistema di microservizi a livello enterpriseGiulio Roggero
 
GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...
GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...
GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...James Anderson
 
Modern Data Architectures for Business Insights at Scale
Modern Data Architectures for Business Insights at Scale Modern Data Architectures for Business Insights at Scale
Modern Data Architectures for Business Insights at Scale Amazon Web Services
 
Machine Learning at Scale with MLflow and Apache Spark
Machine Learning at Scale with MLflow and Apache SparkMachine Learning at Scale with MLflow and Apache Spark
Machine Learning at Scale with MLflow and Apache SparkDatabricks
 
ANIn Chennai April 2024 |Agile Engineering: Modernizing Legacy Systems by Ana...
ANIn Chennai April 2024 |Agile Engineering: Modernizing Legacy Systems by Ana...ANIn Chennai April 2024 |Agile Engineering: Modernizing Legacy Systems by Ana...
ANIn Chennai April 2024 |Agile Engineering: Modernizing Legacy Systems by Ana...AgileNetwork
 
AWS Summit Seoul 2015 - AWS 최신 서비스 살펴보기 - Aurora, Lambda, EFS, Machine Learn...
AWS Summit Seoul 2015 -  AWS 최신 서비스 살펴보기 - Aurora, Lambda, EFS, Machine Learn...AWS Summit Seoul 2015 -  AWS 최신 서비스 살펴보기 - Aurora, Lambda, EFS, Machine Learn...
AWS Summit Seoul 2015 - AWS 최신 서비스 살펴보기 - Aurora, Lambda, EFS, Machine Learn...Amazon Web Services Korea
 
Processing Real-Time Data at Scale: A streaming platform as a central nervous...
Processing Real-Time Data at Scale: A streaming platform as a central nervous...Processing Real-Time Data at Scale: A streaming platform as a central nervous...
Processing Real-Time Data at Scale: A streaming platform as a central nervous...confluent
 
Inawsidom - Data Journey
Inawsidom - Data JourneyInawsidom - Data Journey
Inawsidom - Data JourneyPhilipBasford
 
Ibm Cognos B Iund Pmfj
Ibm Cognos B Iund PmfjIbm Cognos B Iund Pmfj
Ibm Cognos B Iund PmfjFriedel Jonker
 

Similar to MLOps journey at Swisscom: AI Use Cases, Architecture and Future Vision (20)

5 Years Of Building SaaS On AWS
5 Years Of Building SaaS On AWS5 Years Of Building SaaS On AWS
5 Years Of Building SaaS On AWS
 
API and Big Data Solution Patterns
API and Big Data Solution Patterns API and Big Data Solution Patterns
API and Big Data Solution Patterns
 
Achieving Massive Concurrency & Sub-second Query Latency on Cloud Warehouses ...
Achieving Massive Concurrency & Sub-second Query Latency on Cloud Warehouses ...Achieving Massive Concurrency & Sub-second Query Latency on Cloud Warehouses ...
Achieving Massive Concurrency & Sub-second Query Latency on Cloud Warehouses ...
 
Analysing Data in Real-time
Analysing Data in Real-timeAnalysing Data in Real-time
Analysing Data in Real-time
 
DataLive conference in Geneva 2018 - Bringing AI to the Data
DataLive conference in Geneva 2018 - Bringing AI to the DataDataLive conference in Geneva 2018 - Bringing AI to the Data
DataLive conference in Geneva 2018 - Bringing AI to the Data
 
Big Data Analytics Platforms by KTH and RISE SICS
Big Data Analytics Platforms by KTH and RISE SICSBig Data Analytics Platforms by KTH and RISE SICS
Big Data Analytics Platforms by KTH and RISE SICS
 
Blueprint Series: Architecture Patterns for Implementing Serverless Microserv...
Blueprint Series: Architecture Patterns for Implementing Serverless Microserv...Blueprint Series: Architecture Patterns for Implementing Serverless Microserv...
Blueprint Series: Architecture Patterns for Implementing Serverless Microserv...
 
MindSphere: The cloud-based, open IoT operating system. Damiano Manocchia
MindSphere: The cloud-based, open IoT operating system. Damiano ManocchiaMindSphere: The cloud-based, open IoT operating system. Damiano Manocchia
MindSphere: The cloud-based, open IoT operating system. Damiano Manocchia
 
Von der Zustandsüberwachung zur vorausschauenden Wartung
Von der Zustandsüberwachung zur vorausschauenden WartungVon der Zustandsüberwachung zur vorausschauenden Wartung
Von der Zustandsüberwachung zur vorausschauenden Wartung
 
Feature Store as a Data Foundation for Machine Learning
Feature Store as a Data Foundation for Machine LearningFeature Store as a Data Foundation for Machine Learning
Feature Store as a Data Foundation for Machine Learning
 
Platform governance, gestire un ecosistema di microservizi a livello enterprise
Platform governance, gestire un ecosistema di microservizi a livello enterprisePlatform governance, gestire un ecosistema di microservizi a livello enterprise
Platform governance, gestire un ecosistema di microservizi a livello enterprise
 
GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...
GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...
GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...
 
Modern Data Architectures for Business Insights at Scale
Modern Data Architectures for Business Insights at Scale Modern Data Architectures for Business Insights at Scale
Modern Data Architectures for Business Insights at Scale
 
Analytics&IoT
Analytics&IoTAnalytics&IoT
Analytics&IoT
 
Machine Learning at Scale with MLflow and Apache Spark
Machine Learning at Scale with MLflow and Apache SparkMachine Learning at Scale with MLflow and Apache Spark
Machine Learning at Scale with MLflow and Apache Spark
 
ANIn Chennai April 2024 |Agile Engineering: Modernizing Legacy Systems by Ana...
ANIn Chennai April 2024 |Agile Engineering: Modernizing Legacy Systems by Ana...ANIn Chennai April 2024 |Agile Engineering: Modernizing Legacy Systems by Ana...
ANIn Chennai April 2024 |Agile Engineering: Modernizing Legacy Systems by Ana...
 
AWS Summit Seoul 2015 - AWS 최신 서비스 살펴보기 - Aurora, Lambda, EFS, Machine Learn...
AWS Summit Seoul 2015 -  AWS 최신 서비스 살펴보기 - Aurora, Lambda, EFS, Machine Learn...AWS Summit Seoul 2015 -  AWS 최신 서비스 살펴보기 - Aurora, Lambda, EFS, Machine Learn...
AWS Summit Seoul 2015 - AWS 최신 서비스 살펴보기 - Aurora, Lambda, EFS, Machine Learn...
 
Processing Real-Time Data at Scale: A streaming platform as a central nervous...
Processing Real-Time Data at Scale: A streaming platform as a central nervous...Processing Real-Time Data at Scale: A streaming platform as a central nervous...
Processing Real-Time Data at Scale: A streaming platform as a central nervous...
 
Inawsidom - Data Journey
Inawsidom - Data JourneyInawsidom - Data Journey
Inawsidom - Data Journey
 
Ibm Cognos B Iund Pmfj
Ibm Cognos B Iund PmfjIbm Cognos B Iund Pmfj
Ibm Cognos B Iund Pmfj
 

More from BATbern

BATbern52 Moderation Berner Architekten Treffen zu Data Mesh
BATbern52 Moderation Berner Architekten Treffen zu Data MeshBATbern52 Moderation Berner Architekten Treffen zu Data Mesh
BATbern52 Moderation Berner Architekten Treffen zu Data MeshBATbern
 
BATbern52 Swisscom's Journey into Data Mesh
BATbern52 Swisscom's Journey into Data MeshBATbern52 Swisscom's Journey into Data Mesh
BATbern52 Swisscom's Journey into Data MeshBATbern
 
BATbern52 SBB zu Data Products und Knacknüsse
BATbern52 SBB zu Data Products und KnacknüsseBATbern52 SBB zu Data Products und Knacknüsse
BATbern52 SBB zu Data Products und KnacknüsseBATbern
 
BATbern52 Mobiliar zu Skalierte Datenprodukte mit Data Mesh
BATbern52 Mobiliar zu Skalierte Datenprodukte mit Data MeshBATbern52 Mobiliar zu Skalierte Datenprodukte mit Data Mesh
BATbern52 Mobiliar zu Skalierte Datenprodukte mit Data MeshBATbern
 
BATbern52 InnoQ on Data Mesh 2019 2023 2024++
BATbern52 InnoQ on Data Mesh 2019 2023 2024++BATbern52 InnoQ on Data Mesh 2019 2023 2024++
BATbern52 InnoQ on Data Mesh 2019 2023 2024++BATbern
 
Embracing Serverless: reengineering a real-estate digital marketplace
Embracing Serverless: reengineering a real-estate digital marketplaceEmbracing Serverless: reengineering a real-estate digital marketplace
Embracing Serverless: reengineering a real-estate digital marketplaceBATbern
 
Serverless und Event-Driven Architecture
Serverless und Event-Driven ArchitectureServerless und Event-Driven Architecture
Serverless und Event-Driven ArchitectureBATbern
 
Serverless Dev(Ops) in der Praxis
Serverless Dev(Ops) in der PraxisServerless Dev(Ops) in der Praxis
Serverless Dev(Ops) in der PraxisBATbern
 
Serverless at Lifestage
Serverless at LifestageServerless at Lifestage
Serverless at LifestageBATbern
 
Keynote Gregor Hohpe - Serverless Architectures
Keynote Gregor Hohpe - Serverless ArchitecturesKeynote Gregor Hohpe - Serverless Architectures
Keynote Gregor Hohpe - Serverless ArchitecturesBATbern
 
BATbern51 Serverless?!
BATbern51 Serverless?!BATbern51 Serverless?!
BATbern51 Serverless?!BATbern
 
Ein Rückblick anlässlich des 50. BAT aus Sicht eines treuen Partners
Ein Rückblick anlässlich des 50. BAT aus Sicht eines treuen PartnersEin Rückblick anlässlich des 50. BAT aus Sicht eines treuen Partners
Ein Rückblick anlässlich des 50. BAT aus Sicht eines treuen PartnersBATbern
 
From Ideation to Production in 7 days: The Scoring Factory at Raiffeisen
From Ideation to Production in 7 days: The Scoring Factory at RaiffeisenFrom Ideation to Production in 7 days: The Scoring Factory at Raiffeisen
From Ideation to Production in 7 days: The Scoring Factory at RaiffeisenBATbern
 
The Future of Coaching in Sport with AI/ML
The Future of Coaching in Sport with AI/MLThe Future of Coaching in Sport with AI/ML
The Future of Coaching in Sport with AI/MLBATbern
 
BATbern48_ZeroTrust-Konzept und Realität.pdf
BATbern48_ZeroTrust-Konzept und Realität.pdfBATbern48_ZeroTrust-Konzept und Realität.pdf
BATbern48_ZeroTrust-Konzept und Realität.pdfBATbern
 
BATbern48_How Zero Trust can help your organisation keep safe.pdf
BATbern48_How Zero Trust can help your organisation keep safe.pdfBATbern48_How Zero Trust can help your organisation keep safe.pdf
BATbern48_How Zero Trust can help your organisation keep safe.pdfBATbern
 
BATbern48_Zero Trust Architektur des ISC-EJPD.pdf
BATbern48_Zero Trust Architektur des ISC-EJPD.pdfBATbern48_Zero Trust Architektur des ISC-EJPD.pdf
BATbern48_Zero Trust Architektur des ISC-EJPD.pdfBATbern
 
Why did the shift-left end up in the cloud for Bank Julius Baer?
Why did the shift-left end up in the cloud for Bank Julius Baer?Why did the shift-left end up in the cloud for Bank Julius Baer?
Why did the shift-left end up in the cloud for Bank Julius Baer?BATbern
 
Creating a Product through DevOps: The Story of APPUiO Cloud
Creating a Product through DevOps: The Story of APPUiO CloudCreating a Product through DevOps: The Story of APPUiO Cloud
Creating a Product through DevOps: The Story of APPUiO CloudBATbern
 
Zeitnahe Reaktion auf Verordnungsänderungen mit Feature Toggles
Zeitnahe Reaktion auf Verordnungsänderungen mit Feature TogglesZeitnahe Reaktion auf Verordnungsänderungen mit Feature Toggles
Zeitnahe Reaktion auf Verordnungsänderungen mit Feature TogglesBATbern
 

More from BATbern (20)

BATbern52 Moderation Berner Architekten Treffen zu Data Mesh
BATbern52 Moderation Berner Architekten Treffen zu Data MeshBATbern52 Moderation Berner Architekten Treffen zu Data Mesh
BATbern52 Moderation Berner Architekten Treffen zu Data Mesh
 
BATbern52 Swisscom's Journey into Data Mesh
BATbern52 Swisscom's Journey into Data MeshBATbern52 Swisscom's Journey into Data Mesh
BATbern52 Swisscom's Journey into Data Mesh
 
BATbern52 SBB zu Data Products und Knacknüsse
BATbern52 SBB zu Data Products und KnacknüsseBATbern52 SBB zu Data Products und Knacknüsse
BATbern52 SBB zu Data Products und Knacknüsse
 
BATbern52 Mobiliar zu Skalierte Datenprodukte mit Data Mesh
BATbern52 Mobiliar zu Skalierte Datenprodukte mit Data MeshBATbern52 Mobiliar zu Skalierte Datenprodukte mit Data Mesh
BATbern52 Mobiliar zu Skalierte Datenprodukte mit Data Mesh
 
BATbern52 InnoQ on Data Mesh 2019 2023 2024++
BATbern52 InnoQ on Data Mesh 2019 2023 2024++BATbern52 InnoQ on Data Mesh 2019 2023 2024++
BATbern52 InnoQ on Data Mesh 2019 2023 2024++
 
Embracing Serverless: reengineering a real-estate digital marketplace
Embracing Serverless: reengineering a real-estate digital marketplaceEmbracing Serverless: reengineering a real-estate digital marketplace
Embracing Serverless: reengineering a real-estate digital marketplace
 
Serverless und Event-Driven Architecture
Serverless und Event-Driven ArchitectureServerless und Event-Driven Architecture
Serverless und Event-Driven Architecture
 
Serverless Dev(Ops) in der Praxis
Serverless Dev(Ops) in der PraxisServerless Dev(Ops) in der Praxis
Serverless Dev(Ops) in der Praxis
 
Serverless at Lifestage
Serverless at LifestageServerless at Lifestage
Serverless at Lifestage
 
Keynote Gregor Hohpe - Serverless Architectures
Keynote Gregor Hohpe - Serverless ArchitecturesKeynote Gregor Hohpe - Serverless Architectures
Keynote Gregor Hohpe - Serverless Architectures
 
BATbern51 Serverless?!
BATbern51 Serverless?!BATbern51 Serverless?!
BATbern51 Serverless?!
 
Ein Rückblick anlässlich des 50. BAT aus Sicht eines treuen Partners
Ein Rückblick anlässlich des 50. BAT aus Sicht eines treuen PartnersEin Rückblick anlässlich des 50. BAT aus Sicht eines treuen Partners
Ein Rückblick anlässlich des 50. BAT aus Sicht eines treuen Partners
 
From Ideation to Production in 7 days: The Scoring Factory at Raiffeisen
From Ideation to Production in 7 days: The Scoring Factory at RaiffeisenFrom Ideation to Production in 7 days: The Scoring Factory at Raiffeisen
From Ideation to Production in 7 days: The Scoring Factory at Raiffeisen
 
The Future of Coaching in Sport with AI/ML
The Future of Coaching in Sport with AI/MLThe Future of Coaching in Sport with AI/ML
The Future of Coaching in Sport with AI/ML
 
BATbern48_ZeroTrust-Konzept und Realität.pdf
BATbern48_ZeroTrust-Konzept und Realität.pdfBATbern48_ZeroTrust-Konzept und Realität.pdf
BATbern48_ZeroTrust-Konzept und Realität.pdf
 
BATbern48_How Zero Trust can help your organisation keep safe.pdf
BATbern48_How Zero Trust can help your organisation keep safe.pdfBATbern48_How Zero Trust can help your organisation keep safe.pdf
BATbern48_How Zero Trust can help your organisation keep safe.pdf
 
BATbern48_Zero Trust Architektur des ISC-EJPD.pdf
BATbern48_Zero Trust Architektur des ISC-EJPD.pdfBATbern48_Zero Trust Architektur des ISC-EJPD.pdf
BATbern48_Zero Trust Architektur des ISC-EJPD.pdf
 
Why did the shift-left end up in the cloud for Bank Julius Baer?
Why did the shift-left end up in the cloud for Bank Julius Baer?Why did the shift-left end up in the cloud for Bank Julius Baer?
Why did the shift-left end up in the cloud for Bank Julius Baer?
 
Creating a Product through DevOps: The Story of APPUiO Cloud
Creating a Product through DevOps: The Story of APPUiO CloudCreating a Product through DevOps: The Story of APPUiO Cloud
Creating a Product through DevOps: The Story of APPUiO Cloud
 
Zeitnahe Reaktion auf Verordnungsänderungen mit Feature Toggles
Zeitnahe Reaktion auf Verordnungsänderungen mit Feature TogglesZeitnahe Reaktion auf Verordnungsänderungen mit Feature Toggles
Zeitnahe Reaktion auf Verordnungsänderungen mit Feature Toggles
 

Recently uploaded

Effective Strategies for Wix's Scaling challenges - GeeCon
Effective Strategies for Wix's Scaling challenges - GeeConEffective Strategies for Wix's Scaling challenges - GeeCon
Effective Strategies for Wix's Scaling challenges - GeeConNatan Silnitsky
 
WSO2Con2024 - Simplified Integration: Unveiling the Latest Features in WSO2 L...
WSO2Con2024 - Simplified Integration: Unveiling the Latest Features in WSO2 L...WSO2Con2024 - Simplified Integration: Unveiling the Latest Features in WSO2 L...
WSO2Con2024 - Simplified Integration: Unveiling the Latest Features in WSO2 L...WSO2
 
WSO2Con2024 - GitOps in Action: Navigating Application Deployment in the Plat...
WSO2Con2024 - GitOps in Action: Navigating Application Deployment in the Plat...WSO2Con2024 - GitOps in Action: Navigating Application Deployment in the Plat...
WSO2Con2024 - GitOps in Action: Navigating Application Deployment in the Plat...WSO2
 
WSO2CON 2024 - OSU & WSO2: A Decade Journey in Integration & Innovation
WSO2CON 2024 - OSU & WSO2: A Decade Journey in Integration & InnovationWSO2CON 2024 - OSU & WSO2: A Decade Journey in Integration & Innovation
WSO2CON 2024 - OSU & WSO2: A Decade Journey in Integration & InnovationWSO2
 
WSO2CON 2024 - Designing Event-Driven Enterprises: Stories of Transformation
WSO2CON 2024 - Designing Event-Driven Enterprises: Stories of TransformationWSO2CON 2024 - Designing Event-Driven Enterprises: Stories of Transformation
WSO2CON 2024 - Designing Event-Driven Enterprises: Stories of TransformationWSO2
 
From Theory to Practice: Utilizing SpiraPlan's REST API
From Theory to Practice: Utilizing SpiraPlan's REST APIFrom Theory to Practice: Utilizing SpiraPlan's REST API
From Theory to Practice: Utilizing SpiraPlan's REST APIInflectra
 
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...Bert Jan Schrijver
 
WSO2CON 2024 Slides - Unlocking Value with AI
WSO2CON 2024 Slides - Unlocking Value with AIWSO2CON 2024 Slides - Unlocking Value with AI
WSO2CON 2024 Slides - Unlocking Value with AIWSO2
 
WSO2Con2024 - Software Delivery in Hybrid Environments
WSO2Con2024 - Software Delivery in Hybrid EnvironmentsWSO2Con2024 - Software Delivery in Hybrid Environments
WSO2Con2024 - Software Delivery in Hybrid EnvironmentsWSO2
 
AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf
AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdfAzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf
AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdfryanfarris8
 
WSO2Con2024 - Facilitating Broadband Switching Services for UK Telecoms Provi...
WSO2Con2024 - Facilitating Broadband Switching Services for UK Telecoms Provi...WSO2Con2024 - Facilitating Broadband Switching Services for UK Telecoms Provi...
WSO2Con2024 - Facilitating Broadband Switching Services for UK Telecoms Provi...WSO2
 
WSO2CON 2024 - IoT Needs CIAM: The Importance of Centralized IAM in a Growing...
WSO2CON 2024 - IoT Needs CIAM: The Importance of Centralized IAM in a Growing...WSO2CON 2024 - IoT Needs CIAM: The Importance of Centralized IAM in a Growing...
WSO2CON 2024 - IoT Needs CIAM: The Importance of Centralized IAM in a Growing...WSO2
 
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024VictoriaMetrics
 
Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...
Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...
Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...Lisi Hocke
 
WSO2Con2024 - Unleashing the Financial Potential of 13 Million People
WSO2Con2024 - Unleashing the Financial Potential of 13 Million PeopleWSO2Con2024 - Unleashing the Financial Potential of 13 Million People
WSO2Con2024 - Unleashing the Financial Potential of 13 Million PeopleWSO2
 
WSO2CON 2024 Slides - Open Source to SaaS
WSO2CON 2024 Slides - Open Source to SaaSWSO2CON 2024 Slides - Open Source to SaaS
WSO2CON 2024 Slides - Open Source to SaaSWSO2
 
WSO2CON 2024 - How to Run a Security Program
WSO2CON 2024 - How to Run a Security ProgramWSO2CON 2024 - How to Run a Security Program
WSO2CON 2024 - How to Run a Security ProgramWSO2
 
[GeeCON2024] How I learned to stop worrying and love the dark silicon apocalypse
[GeeCON2024] How I learned to stop worrying and love the dark silicon apocalypse[GeeCON2024] How I learned to stop worrying and love the dark silicon apocalypse
[GeeCON2024] How I learned to stop worrying and love the dark silicon apocalypseTomasz Kowalczewski
 
WSO2Con2024 - Navigating the Digital Landscape: Transforming Healthcare with ...
WSO2Con2024 - Navigating the Digital Landscape: Transforming Healthcare with ...WSO2Con2024 - Navigating the Digital Landscape: Transforming Healthcare with ...
WSO2Con2024 - Navigating the Digital Landscape: Transforming Healthcare with ...WSO2
 
WSO2CON 2024 - Unlocking the Identity: Embracing CIAM 2.0 for a Competitive A...
WSO2CON 2024 - Unlocking the Identity: Embracing CIAM 2.0 for a Competitive A...WSO2CON 2024 - Unlocking the Identity: Embracing CIAM 2.0 for a Competitive A...
WSO2CON 2024 - Unlocking the Identity: Embracing CIAM 2.0 for a Competitive A...WSO2
 

Recently uploaded (20)

Effective Strategies for Wix's Scaling challenges - GeeCon
Effective Strategies for Wix's Scaling challenges - GeeConEffective Strategies for Wix's Scaling challenges - GeeCon
Effective Strategies for Wix's Scaling challenges - GeeCon
 
WSO2Con2024 - Simplified Integration: Unveiling the Latest Features in WSO2 L...
WSO2Con2024 - Simplified Integration: Unveiling the Latest Features in WSO2 L...WSO2Con2024 - Simplified Integration: Unveiling the Latest Features in WSO2 L...
WSO2Con2024 - Simplified Integration: Unveiling the Latest Features in WSO2 L...
 
WSO2Con2024 - GitOps in Action: Navigating Application Deployment in the Plat...
WSO2Con2024 - GitOps in Action: Navigating Application Deployment in the Plat...WSO2Con2024 - GitOps in Action: Navigating Application Deployment in the Plat...
WSO2Con2024 - GitOps in Action: Navigating Application Deployment in the Plat...
 
WSO2CON 2024 - OSU & WSO2: A Decade Journey in Integration & Innovation
WSO2CON 2024 - OSU & WSO2: A Decade Journey in Integration & InnovationWSO2CON 2024 - OSU & WSO2: A Decade Journey in Integration & Innovation
WSO2CON 2024 - OSU & WSO2: A Decade Journey in Integration & Innovation
 
WSO2CON 2024 - Designing Event-Driven Enterprises: Stories of Transformation
WSO2CON 2024 - Designing Event-Driven Enterprises: Stories of TransformationWSO2CON 2024 - Designing Event-Driven Enterprises: Stories of Transformation
WSO2CON 2024 - Designing Event-Driven Enterprises: Stories of Transformation
 
From Theory to Practice: Utilizing SpiraPlan's REST API
From Theory to Practice: Utilizing SpiraPlan's REST APIFrom Theory to Practice: Utilizing SpiraPlan's REST API
From Theory to Practice: Utilizing SpiraPlan's REST API
 
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
 
WSO2CON 2024 Slides - Unlocking Value with AI
WSO2CON 2024 Slides - Unlocking Value with AIWSO2CON 2024 Slides - Unlocking Value with AI
WSO2CON 2024 Slides - Unlocking Value with AI
 
WSO2Con2024 - Software Delivery in Hybrid Environments
WSO2Con2024 - Software Delivery in Hybrid EnvironmentsWSO2Con2024 - Software Delivery in Hybrid Environments
WSO2Con2024 - Software Delivery in Hybrid Environments
 
AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf
AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdfAzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf
AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf
 
WSO2Con2024 - Facilitating Broadband Switching Services for UK Telecoms Provi...
WSO2Con2024 - Facilitating Broadband Switching Services for UK Telecoms Provi...WSO2Con2024 - Facilitating Broadband Switching Services for UK Telecoms Provi...
WSO2Con2024 - Facilitating Broadband Switching Services for UK Telecoms Provi...
 
WSO2CON 2024 - IoT Needs CIAM: The Importance of Centralized IAM in a Growing...
WSO2CON 2024 - IoT Needs CIAM: The Importance of Centralized IAM in a Growing...WSO2CON 2024 - IoT Needs CIAM: The Importance of Centralized IAM in a Growing...
WSO2CON 2024 - IoT Needs CIAM: The Importance of Centralized IAM in a Growing...
 
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
 
Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...
Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...
Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...
 
WSO2Con2024 - Unleashing the Financial Potential of 13 Million People
WSO2Con2024 - Unleashing the Financial Potential of 13 Million PeopleWSO2Con2024 - Unleashing the Financial Potential of 13 Million People
WSO2Con2024 - Unleashing the Financial Potential of 13 Million People
 
WSO2CON 2024 Slides - Open Source to SaaS
WSO2CON 2024 Slides - Open Source to SaaSWSO2CON 2024 Slides - Open Source to SaaS
WSO2CON 2024 Slides - Open Source to SaaS
 
WSO2CON 2024 - How to Run a Security Program
WSO2CON 2024 - How to Run a Security ProgramWSO2CON 2024 - How to Run a Security Program
WSO2CON 2024 - How to Run a Security Program
 
[GeeCON2024] How I learned to stop worrying and love the dark silicon apocalypse
[GeeCON2024] How I learned to stop worrying and love the dark silicon apocalypse[GeeCON2024] How I learned to stop worrying and love the dark silicon apocalypse
[GeeCON2024] How I learned to stop worrying and love the dark silicon apocalypse
 
WSO2Con2024 - Navigating the Digital Landscape: Transforming Healthcare with ...
WSO2Con2024 - Navigating the Digital Landscape: Transforming Healthcare with ...WSO2Con2024 - Navigating the Digital Landscape: Transforming Healthcare with ...
WSO2Con2024 - Navigating the Digital Landscape: Transforming Healthcare with ...
 
WSO2CON 2024 - Unlocking the Identity: Embracing CIAM 2.0 for a Competitive A...
WSO2CON 2024 - Unlocking the Identity: Embracing CIAM 2.0 for a Competitive A...WSO2CON 2024 - Unlocking the Identity: Embracing CIAM 2.0 for a Competitive A...
WSO2CON 2024 - Unlocking the Identity: Embracing CIAM 2.0 for a Competitive A...
 

MLOps journey at Swisscom: AI Use Cases, Architecture and Future Vision

  • 1. MLOps Journey at Swisscom AIUseCases,ArchitectureandFutureVision Joana Soares Machado Maxime Darçot 24.02.2023
  • 2. Outline 2 • Overview: AI Use Cases at Swisscom • Challenges • AI Use Cases Deep Dive • MLOps Principles • Future Outlook: The AWS Move MLOps journey at Swisscom
  • 3. Overview AI Use Cases at Swisscom 3
  • 4. Infrastructure Analytics Anomaly Detection, Network Optimization, Customer Centric Monitoring, Mobility Insights Conversational AI Conversational AI for Swisscom’s Products & Services, Innovation & Research Data Services Data Lake & Warehouse, Enablers, Data and Analytics Tools in Self-service B2B Analytics Business Insights about B2B customers and Process Performance, KPIs definition Mass Market Analytics SC Touchpoints Insights, Pricing Simulations for Offers, Recommender and Offer System to Sales Data, Analytics & AI Business Analytics Analytics solutions for Sales, Logistics, Finance, Controlling and HCM, Management Reporting 4 MLOps journey at Swisscom
  • 5. 5 350 experts in Data, Analytics & AI Skill Overview • Data Science / AI Solution Design • Software Development • Data Engineering • Natural Language Processing • Business Engineering & Design • Reporting & Analysis Specialists • User Experience Design Data & Infrastructure Overview • >660 servers' on-premise infrastructure • >40 Data, Analytics and AI services & platforms & > 150 tools • 10.73 Peta Bytes (PB) data • 8.17 Mio. real-time messages per second • 4’756 monthly active users of dashboard tools MLOps journey at Swisscom
  • 7. 7 MLOps journey at Swisscom Data Regulations CH-only, on-prem Data Formats Tabular, time series Challenges: Diversity of... Scalability Req. Nb of models, SLAs • Millions of models @INI (due to anomaly detection) • 100s of models (@B2C & @B2B) • Customer facing applications Tech Stacks Clouds, legacy • k8s on AWS & SBD • Internal Application Cloud (CF) AI Applications NLP, ASR, Real-time • 40 product instances @DNA • 350 Users @DNA • ML Engineers • Business Analysts • 3000 Users @Swisscom
  • 9. 9 Digital Marketplace 35M API calls/day Software Applications 300M Function calls/day Business Processes 80K Processes/day Network Infrastructure 20B Network interactions/day MLOps journey at Swisscom Anomaly Detection Use Cases at Swisscom Source: Towards Data Science Time series Anomaly Detection
  • 10. 10 Data Sourcing (Time series) Real-time Anomaly Detection UI + Alerting Outage Detection Insights Feedback MLOps journey at Swisscom Anomaly Detection Pipeline / Daily Data Preparation Training + Batch Prediction
  • 11. 11 MLOps journey at Swisscom Anomaly Detection Pipeline: Open Source Tools CI/CD Version Control System Workflow Orchestration ML Resource Infrastructure ML Metadata Store ML Model Serving Pipeline Monitoring
  • 12. • Measure lags and pipeline availability. • Different aspects of quality: − Training data validation; − Deploying models with good accuracy; − Logging information about models and errors, for transparency and reproducibility. • SLI/SLO alerting: target over time period. 12 MLOps journey at Swisscom MLOps: Managing the ML Lifecycle Source: Sensu
  • 15. 15 MLOps journey at Swisscom Monitoring over time – Requests and Processing Time
  • 16. 16 MLOps journey at Swisscom Conversational AI – Dialog flow 16 Language: EN Intent: Customer has a technical issue Topics: Internet Hello I see that your internet connection has been blocked because you have 3 outstanding bills Do you wish to get more time to pay these bills? Hello and welcome to Swisscom, how may we help you today? My internet stopped working Yes please Would you be able to pay by the end of next month? Sure, even by the 15th Alright I’ve unblocked your connection, you have until April 15th to pay these 3 bills. Thank you Intent: Customer confirms Topics: None Intent: Customer confirms Topics: None Intent: Customer thanks us Topics: None Language Detection Intent & Topic classifiers General Dialog Manager Billing APIs Intent: Customer greets us Topics: None Background call to billing API -> Customer is blocked Entity Recognition Entity: 15th (Date) Domain classifier Domain: None Domain: Technical support Domain: Billing Technical Support Dialog Manager Billing Dialog Manager General Dialog Manager Background call to billing API -> Change due dates
  • 17. 17 Speech Recognition Language Detection Entity Recognition Machine Translation Intent & Topic classifiers Retrieval System Domain classifier … Monitoring Reporting Model Storage CI / CD Error alerting Secure data storage Flow design tool Data labeling tools End to end testing tool Training pipelines Billing APIs CRM … Hotline Whatsapp TV Box E-mail Letter / Fax ... Input APIs Dialog Managers Conversational AI stack
  • 18. 18 Input Channels AI/ML layer Runtime foundations layer Offline tools layer Input APIs Integration layer Dialog Managers Conversational AI stack
  • 19. 19 Speech Recognition Language Detection Entity Recognition Machine Translation Intent & Topic classifiers Retrieval System Domain classifier … Monitoring Reporting Model Storage CI / CD Error alerting Secure data storage Flow design tool Data labeling tools End to end testing tool Training pipelines Billing APIs CRM … Hotline Whatsapp TV Box E-mail Letter / Fax ... Input APIs Dialog Managers Conversational AI – ML Operations
  • 20. 20 Intent & Topic classifiers Domain classifier Model Storage CI / CD Secure data storage Data labeling tools Training pipelines Conversational AI – ML Operations Intent & Topic classifiers Domain classifier Model Storage CI / CD Secure data storage Data labeling tools Training pipelines
  • 21. 21 Intent & Topic classifiers Domain classifier Model Storage CI / CD Secure data storage Data labelling tools Training pipelines Conversational AI – ML Operations
  • 22. 22 MLOps journey at Swisscom Conversational AI: MLOps Tools CI/CD Version Control System Workflow Orchestration ML Resource Infrastructure Secure Data Storage AI Labelling Tool ML Model Serving
  • 24. 24 (ML) Product Lifecycle SW Product 1. Development 2. Integration 3. Testing 4. Deployment 5. Feedback 6. Monitoring 7. Operations ML Product 2. Data Preparation 1. Data Collection 3. Data Integration 4. Data Transformation 5. Model Training 6. Model Registration 7. Model Deployment 8. Model Evaluation 9. Model Monitoring A. Data Engineering B. Model Engineering C. Model Operations MLOps journey at Swisscom
  • 25. 25 Tools for the MLOps Principles over the Data Product Lifecycle Principle Collaboration [P5] Reproducibility [P3] CI/CD automation [P1] Workflow orchestration [P2] Continuous ML training/evaluation [P6] Feedback loops [P9] ML metadata tracking/logging [P7] Continuous monitoring [P8] Versioning of data, code, model [P4] Adapted from: Machine Learning Operations (MLOps): Overview, Definition, and Architecture C. Model Operations A. Data Engineering B. Model Engineering Version Control System P4, P5 CI/CD Component P1, P6, P9 ML Resource Infrastructure P6 Feature Store System P3, P4 Workflow Orchestration Component P2, P3, P6 ML Model Registry P3, P4 Monitoring Component P8, P9 ML Metadata Store P4, P7 ML Model Serving Component P1 MLOps journey at Swisscom
  • 26. MLOps journey at Swisscom 26 MLOps Harmonization – A Central Solution at Swisscom? Standardized solution Easy to start new project No need to reinvent the wheel every time Collaborate and exchange best practices
  • 27. • How much of the requirements does it cover? • How good is the user experience? • What is outlook on future requirements? • What is outlook on future UX ? 27 MLOps Solutions - Evaluation Dimensions • Development & Integration Costs. • Operations & Maintenance Costs. • Infrastructure (incl. Licenses) Costs. • Outlook on cost evolution. User Needs & UX Costs & Efficiency MLOps journey at Swisscom Personas / Roles • Yes-Code vs No-Code: − Data Engineer, Data Scientist, (ML)Ops Engineer, SW Engineer. − Business Analyst, Product Manager.
  • 29. 29 We want to converge our warehouses & Big Data lake into a hybrid lake house architecture From on prem only Data Lake & Warehouse Warehouses Swisscom Data Warehouse ONDS B2B DWH Swisscom Big Data Platform Data Lake Hybrid Data Factory + To a hybrid Data Lakehouse One Data Platform MLOps journey at Swisscom
  • 30. 30 One Data Platform – High-level Architecture Store & Process Applications & Data Functions Exposure Gatekeeper Gatekeeper MLOps DWH Analytic s e.g. Time Series Custom Apps e.g. Advanced Processing Ingestion Storage Transformations Ingestion Storage Transformations Data Hub On Premise (for very specific workloads) AWS (whenever possible) Gatekeeper Gatekeeper Data Access DWH Analytics e.g. Time Series MLOps Custom Apps e.g. Advanced Processing real-time / batch real-time / batch CI /CD Data Management Foundation 3 2 1 4 6 5 7 8 9 4 MLOps journey at Swisscom
  • 31. • AWS Sagemaker • All-in-one, yet modular • Including no-code option • Amazon MWAA • Managed Airflow (orchestration) • DataRobot • All-in-one • Including no-code option 31 MLOps - Commercial Product Candidates • All-in-one • Databricks, Snowflake • Single components • Fiddler AI (model performance management) • Arize (observability) • CometML (model management & monitoring) • Astronomer (orchestration) Concrete Options Other Possibilities MLOps journey at Swisscom
  • 32. 32 MLOps – Open Source Tools MLOps journey at Swisscom Pros • Free solutions, community built. • Continuously improved and well tested. • Independent solutions for each component: • Fully customizable and modular. • Easy switch when better options available. • Infrastructure agnostic (Hybrid cloud). Cons • Needs to be managed by Swisscom engineers. • May be more costly than “managed”solutions in terms of infrastructure. • Most solutions don’t offer a “no-code”option.
  • 33. 33 MLOps – Hybrid Solution (AWS & Open source) MLOps journey at Swisscom Bringing together the best of both worlds?
  • 34. • Huge diversity of AI use cases. • Need for harmonization in the MLOps area. • The AWS move brings benefits but also questions. • A hybrid setup will likely be a main part of the solution. • This is work in progress, our MLOps journey is not at an end yet. Summary 34 MLOps journey at Swisscom
  • 35. 35 Thank you! Maxime Darçot maxime-darcot Joana Soares Machado joana-soares-machado