How to Innovate Your Software with ML? - Amazon Web Services... · project management scrum skill...
Transcript of How to Innovate Your Software with ML? - Amazon Web Services... · project management scrum skill...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
How to Innovate Your Software with ML?
Dr. Jürgen NützelCEO4FriendsOnly.com Internet Technologies AG
Thomas SchlerethManaging Director Can Do GmbH
Markus OponczewskiDirector Business Unitfme AG
Raul FiruCTOHaufe Group
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
The Lack of skilled Professionals – From Crisis to Opportunity with AIThomas Schlereth, Can Do GmbH
AWS Conversational ML Services and ECMMarkus Oponczewski, fme AG
Behind the Scenes – Haufe’s AI JourneyRaul Firu, Haufe Group
How QuickSight visualizes KPIs of the E-Commerce Recommendation EngineDr. Jürgen Nützel, 4FriendsOnly.com Internet Technologies AG
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
How QuickSight visualizes KPIs of the E-Commerce Recommendation Engine
Dr. Jürgen NützelCEO4FriendsOnly.com Internet Technologies AG
AWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 4
How QuickSight visualizes KPIs of
the E-Commerce
recommendation engine
4FriendsOnly.com Internet Technologies AG– Spin-off of Fraunhofer IDMT & university
– CEO, lecturer and main owner: Dr. Jürgen Nützel
– E-Commerce experts (> 10 years)
– We focus also on
• Mobile development
• Cloud-computing (AWS since 2013)
• Digital transformation
– Amazon AWS partner since 2017
– AWS user group lead
AWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 5
One of our Customer – Papier LIEBLa B2B Office Wholesaler from Regensburg …
AWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 6
We develop and run their
B2B online shops on AWS:
www.liebl.de,
procurement.liebl.de
… asked us for a
recommendation service
which categorizes online
shoppers very fast.
Recommendations in the ShopAWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 7
In the online shop
with active
debugging info
How we get Recommendations• Assumption
Users with similar behaviour (in the shop) have similar interests
• DataUser’s click stream in the shop
Mobile or desktop user
Shopping time (Friday and weekend is different, especially in B2B)
• CalculationUsers similarity matrix. Similar means having same products clicked/bought
The products the most similar users have clicked/bought most
• ImplementationSimilarity matrix will be updated in memory in real-time (using Node.js)
AWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 8
Setup we start with
AWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 9
E-Commerce-Cluster calls
protected API to receive
recommendations
Browser sends click stream data
to the server{"action":"addToCart","data":{"sku":"50189","
price":"3.27","quantity":"5"}}
Drawing made with: https://cloudcraft.co
How QuickSight comes into play
AWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 10
We can pre-configure any
diagram to show the shop
manager all wanted KPIs (not
only recommendation related)
QuickSight is not available in
Frankfurt. We had to
replicate the data to Ireland
We manage the data in QuickSightAWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 11
We prepare analyses and dashboardsAWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 12
AWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 13
Embedding QuickSight Dashboards(using the Enterprise edition)
We show our customer the
success of the
recommendation
Thank you
Follow us on Meetup:
https://www.meetup.com/de-DE/Ilmenau-Amazon-Web-Services-Meetup/
Dr. Jürgen Nützel
CEO
www.4fo.de/en
AWS Summit Berlin 02/2019 Dr. Jürgen Nützel, [email protected], 4FriendsOnly.com Internet Technologies AG 14
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
The Lack of skilled Professionals –From Crisis to Opportunity with AI
Thomas SchlerethManaging Director Can Do GmbH
THE LACK OF skilled professionals
From crisis to opportunity with AI
More than 80,000 IT specialists are
desperately sought in GermanyWAR for Talents
Can Do GmbH Provider of standard software
for professional resource
management and skill-based
resource planning
CustomersUsing Can Do, more than 80
companies - mostly corporate
groups in DACH - benefit from
an optimized deployment of their staff
• Best possible deployment of staff
• Comprehensive risk management
• New quality of resource planning due to skills
• Recommended for action given by Can Do AI
• Company-wide project management &
resource management
PROJECT SUCCESS
thanks to Can Do
project management
scru
m
skill managementbudget management
sprint
eppi
cs
Project Portfolio Management
project score
proj
ect c
osts
bottlneck resource permanent target-actual comparison
jira integration
time recording
demand management
base load
agile organization
Hybrid Project ManagementW
ater
mod
el®
-dy
nam
icca
paci
ty
bala
ncin
gm
etho
d
plan
cos
ts
algorithm: planning with
inaccurate data
wor
kloa
d
skill
cate
gory
interfacesmanagement dashboards
skill based resource management
RANGE OF SERVICE
ARTIFICIAL INTELLIGENCE
vacation planning
approval procedure
capacity
analysis
real
-tim
e co
mp
uti
ng
ai: expert system
cost report
pull-off feature
project progress
performance analysis
baseline analysis
portfolio management
reso
urc
e m
anag
emen
t
epics
overload control
inter- and
Cross project
linkages
Gaussian standard normal distribution
user-centric
app model
role
-spe
cific
user
-frie
ndly
milestones
Gantt chartGantt chart
sub-projects
Pro
babi
lity
of o
ccur
renc
e of
pro
ject
s
Control over approx. 20 risk types
use case oriented
Can
Do
IT TAKES A LOT OF EFFORT AND
MONEY TO MAKE THIS FLY
MASSIVECOMPUTING POWER REQUIRED
1 ressource with 10 inaccurate planned working packages
Example: Planing with inaccurate data:
Inaccuracy Calculation Layer ( Gauss ) & Watermodel® algorithmAPPROX. 500.000 ITERATIONS
CPU: 32 CPU-Cores > 2 GHz
RAM: min. 64 GB
RECOMMENDING AI Learning System for indepth
analysis
Monitoring AI Pattern Matching
AWS
o Trial and Error
o Preparation of a requirement specification:
• Hardware Costs
• Licensing Costs
• Implementation Costs
• Training Costs
• Costs for software & hardware operation
COST & EFFORT COMPARISON
TOTAL USER COSTS
On Site
o Selection process: RFI, RFP, long list, short list
o Preparation of a requirement specification:
• Hardware Costs
• Licensing Costs
• Implementation Costs
• Training Costs
• Costs for software & hardware operation
faster, better, cheaper
ML in AWS is
Thomas Schlereth
Managing Director
Can Do GmbH
thomas.schlereth@
can-do.de
www.can-do.de
Ensuring the best possible deployment of your staff!
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
AWS Conversational ML Services and ECM
Markus OponczewskiDirector Business Unitfme AG
AWS Conversational ML Services and ECMAWS Summit 2019 Berlin – Software Innovation Forum
Markus Oponczewski
@moponczewski
fme – Who we are
BS, MUC, FFM – Danbury (US) – Cluj-Napoca (RO)
250+ Talented People & Dedicated Professionals
Enterprise Content / Digital Transformation / Cloud Computing
Cloud Process Expertise: Cloud-native, DevOps, CloudOps
Cloud Platform Expertise: AWS & Pivotal
What is an ECM System ?
30
Think of…
- Document Management
- Content & Media Handling
- Sensitive Data Protection
- Global Collaboration & Workflow
- Archival Solutions (DSGVO)
- Enterprise Requirements
The Challenge
#6120 © fme AG
Extend System Operability
without changing the System !
Extend System Functionality
without changing the System !
Amazon Alexa
Comprehend, Translate, Polly
Demo Time
32
Watch The Full Videohttps://www.youtube.com/watch?v=jKGAeEI8m5g
© fme AG
Show Me The Folder
Open The Document
Read/ Translate The Document
Read/ Translate The […] Section
Read/ Translate Page Number […]
Translate The Document Into […]
Translate And Save in DCTM
Extend Operability with Alexa & Extend Functionality with Comprehend, Translate and Polly
Interact with
• Alexa
• Comprehend
• Translate
• Polly
The Architecture
33 © fme AG
The Architecture
34 © fme AG
REST-API
-
API
-
SDK
Any Existing or
Legacy System
API Gateway
Other UseCases
Extend Functionality with Machine Learning Services
- Pimp Incident & Ticket System (ITSM)
Extend Operability with Alexa Voice UI
- Chemical & Pharmaceutical Lab environments
#6120 © fme AG
It’s fme you love to work with.
fme AG | Rupprechtstr. 25| D-80636 München | www.fme.de
Markus Oponczewski
@moponczewski
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Behind the Scenes – Haufe’s AI Journey
Raul FiruCTOHaufe Group
BEHIND THE SCENES
The haufe group
Why is ai so different?
μ
Foundations vs. what you see
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
M L F R A M E W O R K S &
I N F R A S T R U C T U R E
A I S E R V I C E S
R E K O G N I T I O N
I M A G E
P O L L Y T R A N S C R I B E T R A N S L A T E C O M P R E H E N D L E XR E K O G N I T I O N
V I D E O
Vis ion Speech Language Chatbots
A M A Z O N
S A G E M A K E R
B U I L D T R A I N
F O R E C A S T
Forecast ing
T E X T R A C T P E R S O N A L I Z E
Recommendat ions
D E P L O Y
Pre-bui l t a lgor i thms & notebooks
Data label ing (G R O U N D T R U T H )
One-c l ick model t ra in ing & tuning
Opt imizat ion (N E O )
One-c l ick deployment & host ing
M L S E R V I C E S
F r a m e w o r k s I n t e r f a c e s I n f r a s t r u c t u r e
E C 2 P 3
& P 3 N
E C 2 C 5 F P G A s G R E E N G R A S S E L A S T I C
I N F E R E N C E
Reinforcement learningAlgor i thms & models ( A W S M A R K E T P L A C E
F O R M A C H I N E L E A R N I N G )
“Starting at the bottom”
μ
Semantics analytics based recommender
“Starting in the middle”Organizational analytics – semantics based clustering
“Starting at the top”NLP Based search for all haufe content
Takeaways
“It is difficult to make predictions, especially
about the future.”
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
AI and Machine Learning ISV Event in Germany
Deep dive workshops
Separate Business and Technical tracks
Timing: May/June 2019
Attractive prizes for selected participants
Networking opportunity
S U M M I TB ERL I N