AWS Construction Event for Gen AI and Connected Data Lakes - Jun 2024

AWS Construction Event for Gen AI and Connected Data Lakes

1. Data and AI in
Real Estate & Construction
© 2 024 Cogniza nt
June 2024

2. : UK&I Consulting
“
Inawisdom was
founded with a
simple goal to give
our customers the
ability to exploit
every aspect of
their data using AI.”
- Robin Meehan,
co-founder & MD
Cognizant's UK&I is based in the UK, with an HQ located in the heart
of the City in London & offices throughout the UK, working
withglobal organisations in the UK, Ireland and across Europe.
Cognizant's UK&I has a team of proven experts and is an Amazon
Web Services (AWS) Premier Services Partner, all-in AWS and
leads the way in full-stack Cloud Engineering, Data Engineering,
Advanced Analytics and AI/ML.
Cognizant believes that combining Inawisdom and Contino with their
wider capabilities and in-depth industry knowledge, it gives our
clients a strategic advisor & partner that they can trust to help them to
embrace and adapt to ever-changing world
Cognizant's UK&I Consulting was relaunched in Jan 2024 following the
acquisitions of Inawisdom (a leadingspecialist in Data, Advanced
Analytics and AI / ML) and Contino (a leadingspecialist in Digital &
Business Transformation). Both have proven & in-depth
experience in delivering outcomes to customers
Who
Why
Where
How

3. Cognizant works at all levels of organizations to
ensure decision makers are given the insights
needed to be able to make decisions using
Data & AI.
Cognizant is able support organizations in
trusting in decisions made with data & AI,
including the ability to share information safey.
Cognizant is able help organizations help
identify their data, understand their data and
help define what their data means order to
power AI Use Cases.
Cognizant aids organizations in coping with the
demands of regulation and the ethical use of
Data & AI.
Literacy & Accessibility
Value & Impact Trust & Maturity
Principles & Compliance
Data & AI
Consulting

4. Current Benefits to Business
ANALYTICS, AI & ML BUSINESS OPPORTUNITIES
Bottom line
benefits
Top line
benefits
New revenue
streams
Increase customer
wallet or market share
Brand perception
or USP
Improve customer acquisition,
conversion and retention
Compliance Better assessment
of risk
Intelligent Process
Automation
Optimisation and
Efficiency Gains

5. Generative AI
Built on the last 30+ Years of progress
Vast “vetted” corpuses are now available
The Cloud has made huge amounts compute power available
via on demand consumption
Advances in AI architecture, especially on attention and
transformers
Simplification of use
“We are at the iPhone moment for AI.”
Jensen Huang, Chief Executive Officer, Nvidia

6. : Full-stack Data Capability
Business
Differentiation/Value
Data Driven
Business Decisions
Cloud Transformation
Adoption and Scale
Digital
Enablement
AI and Machine Learning
Data & Analytics
Data Foundations
Cloud Infrastructure
Landing zone, Control Tower, migration
AI and Machine Learning
Data & Analytics
Data Foundations
Cloud Infrastructure
Landing Zone, Control Tower, migration

7. © 2 024 Cogniza nt7
Facilities Management encompasses a range of
services to ensure the functionality, comfort, safety,
and efficiency of a built environment. It includes
physical security, workspace management, facility
maintenance, and office management
Facilities Management
Use of Generative AI, Advance Analytics & Data
Data and AI can contribute significantly to health and
safety in construction by predicting potential
incidents, optimizing safety measures, and ensuring
compliance through advanced analytics and
machine learning models.
Health & Safety
Data and AI enhance can sustainability in
construction by enabling predictive analytics for
resource optimization, waste reduction, and energy
efficiency, leading to greener building practices.
Sustainability
Building digital twins with data and AI enables real-
time monitoring, predictive maintenance, and
lifecycle management, enhancing infrastructure
management and operational efficiency.
Managing Infrastructure
Data-driven insights to optimize project timelines,
budget management, and resource allocation for
enhanced project delivery.
Project Delivery
Using data to plan future estate plans. This includes
both expansion of existing facilities and identifying
optimal locations to build new facilities and/or
Infrastructure including ESG impact & subsidies
Estate Planning
Leveraging Data and AI to optimize the supply chain
in construction by enhancing visibility (including Data
Sharing), predicting demand, automating
procurement, and improving logistics for efficient
material flow.
Commercial & Supply
Chain
Applying Generative AI helps in matching skills and
people in construction by analysing resumes and job
descriptions, identifying skill gaps, and suggesting
suitable training or upskilling opportunities
Skills & People
Data and AI meeting compliance in construction
including maintaining the 'golden thread' of
information or understanding of changes to
regulation.
Compliance & Regulation

8. Analytics & Insights
Using data-driven insights to optimize project
timelines, budget management, and resource
allocation for enhanced project delivery.
AI & Machine Learning
AI in construction project planning optimises risk
management, and task prioritisation and enhances
decision-making, leading to safer, more efficient, and
cost-effective builds
PROJECT
INSIGHTS
Accessible Data
Making sure that data is easily retrievable, well
managed, understood and usable, fostering informed
decision-making and supporting a data-driven culture
within an organization.
The tracking projects and if they are
running on time and within the
budget.
Spotting early projects that are not
running on time and within the budget,
Suggesting the best remediation action to
be taken.
Identifying faults in design or construction
using Sensors, Vision, COBie, BIM to
perform real-time inspection and analysis
Tracking compliance with Health and
Safety Protocols

9. 99
CASE STUDY
The Customer:
AI in Action: Health and Safety
The Sector: Construction
The Result:
The Solution:
The Requirement:
➢Data Lake using AWS Data Warehouse tool Redshift for construction
projects, containing past project data that includes budget, hours and
health and safety incident data
➢Our exploration of the data found the number of H&S incidents was a
strong indicator of poorly run sites, and poorly run sites historically
incurred higher overspend
➢Machine Learning model was built in Amazon SageMaker (cloud
Machine Learning platform) to predict likely number of incidents
Multinational
infrastructure group
Increase the profitability of projects by predicting
likely overspend
➢Full Productionisation of the Data Lake including ingesting Field360
and unstructured data from Project Portal
➢Data Lakes are now judged to be essential and are used on joint
ventures for some of the largest UK Government infrastructure
projects to provide deep insights

10. GENERATIVE
SEARCH
The ability to search large amounts of
content, i.e.
Advance Search / QA
The ability to search inside private document,
images or websites to find related content and
then returning that content. This typically uses a
LLM like GPT 3.5 but on your own data
Retrieval-Augmented Generation
LLMs are trained on historic records and that
means the results of LLM are from a snapshot
in time. However, RAG allows agents /
integrations with live systems to augment the
results with up-to-date information.
Security & Privacy
Private FMs are not like Internet SaaS Products,
your data is not shared and is kept securely.
This allows for compliance with regulation and
privacy controls.
The ability to search guidelines &
regulations
The ability ask question of previous
projects and gather insights
The ability to describe faults, perform
diagnosis and then return results of
which paths parts & manuals.

11. Text Summarisation
Generates new text that summarises the content
contained from hundreds pages. This is typically used
to pull out the key terms from very verbose
documents
Extraction of Data
Used as part of IDP to extract structured information
from text and images. Examples are invoice line
items or complex nested data points where the
relationship between them holds meaning

Text Comparison
The ability to compare how one document matches
the contents of another
Comparing and updating internal
guidelines according to changes in
regulations
Digitalised items from invoices and
receipts, validation of certificates
Skill Matching and Task Allocation,
Prioritisation and Validation
The ability to help the business
understand what is contained in their
unstructured or semi-structured data
IDP+
Text Classification
The ability to look over the entirety of a piece of
content or document to understand the type or use of
the document
Initial drafting of documents & contacts
with client and suppliers

12. 1212
LIVE - CASE STUDY
The Customer:
IDP - From Document-led to a Data-driven Market-place
The Sector: Insurance
The Solution:
The Result:
The Requirement:
➢Trained and Deployed fine-tuned LLMs targeted at domain specific documents
➢Established an automated, scalable underwriting process to improve
underwriters’ day to day operations and drive business growth
➢Created intelligent AI solution to extract key data points (pricing/policies) from
broker documents held in multiple types (pdf, email, xls)
➢Enabling faster velocity and quality for risk writing, encompassing various
components and personas, to drive profitable business
➢Exploiting new innovations to improve accuracy in rating, forecasting, pricing
and binding risk
➢Reducing operational costs
➢Creating a next-generation of market solutions to enable the business to be
‘future fit’
➢Leading the digital revolution within the underwriting and risk process
International insurance
and reinsurance group
Revolutionise the approach for underwriting risk in specialty
insurance – leveraging AI & automated document processing

13. DATA SHARING
The ability to share data securely and
respond to data access requests
Data Products
A data product is a curated dataset that is self-
service accessible, securely shared, and serves
as a primary source for analytics, documented
with business context, refresh frequency, and
quality details.
Clean Rooms
A data clean room is a secure environment
where data products can be shares and
analysed under strict protocols to maintain
confidentiality and compliance with privacy
regulation
Data Catalog and Lineage
Data Catalog and Lineage involve organizing
and managing data products, ensuring a clear
understanding of data origin, transformation,
and usage, and enhancing trust and governance
across an organization's data ecosystem
Compliance with Golden Thread by
providing access to all artefacts and
contracts from a project
Improved supply chain management
optimisation by sharing data with suppliers
Using 3
rd
Party data sources to improve
insights. Ie. Weather Patterns

14. GIS SIMULATION
The ability to generate and simulate
the ESG impacts that a large project
will have on the real-world
State of The Art Models
Using domain specific machine learning
models created by the Imperial College
London to predict ESG impact
Simulation of Outcomes
The next phase will look to optimise
configurations of real-world resources to
create plausible and optimized layouts for
Carbon reduction
Digital Twin
The creation of a series of Digital Twins
to represent real world objects from
photos take from drones to build a 3D
model.
CARBON SYNC

15. 1515
CASE STUDY
The Customer:
AI in Action: Using AI and ML for Predictive Maintenance
The Sector:
The Result:
The Solution:
The Requirement:Predictive Maintenance service offering - to deliver
innovation and increased value to vehicle owners and operators
ConstructionEquipment
Manufacturing
➢Build of an IoT-enabled Data Lake to ingest sensor data and alerts from
250,000 vehicles across the EMEA and American markets
➢Utilised AWS Control Tower to ensure AWS account management, security
and automation for the customers, using Guardrails for governance
➢Ingested data at three velocities (fast, med, slow) into the common store
➢Transformed the data to an efficient, common format within a single logical
database, with querying and reporting capabilities
➢Driving the evolution of new data-led solutions for vehicle owners, dealers,
manufacturersand operators
➢Created and delivered a roadmap for enhanced service offerings including:
▪Implementation of real-time analytics and alerting to customers
▪Expanding data to include vehicle and devices from the Asian market
▪Manage other devices and manufacturing data sources
▪Implement Machine Learning for accurate predictive maintenance

16. 1616
CASE STUDY
The Customer:
AI in Action: Personal Protective Equipment
The Result:
The Solution:
The Requirement:
➢Used an ADLINK Neon-1040 Camera and inference engine at the site
that runs a custom computer vision model
➢The custom computer vision model is trained in an AWS cloud
Machine Learning platform (SageMaker) to provide the graphics
acceleration compute power needed, and then converted to
OpenVINO to run on device
➢Built on Deep Learning ML (TensorFlow), needing labelling of images
to train the model to detect the PPE using AWS data labelling service
➢The successful detection of:
➢Safety helmet
➢Hi-Vis jacket
➢Person
The Sector: Construction
Multinational
infrastructure group
The ability to detect that the correct PPE is being
worn on a site to reduce risk of H&S incidents

17. 1717
CASE STUDY
The Customer:
The Sector: Steel
Manufacturing industry
The client will be launching the core service to their customers in April 2022.
Further phases are underway to provide valuable insights into production
statistics and error logs from devices through a search capability. Machine
Learning analytics will then provide predictive maintenance alerts.
The Solution:
The Requirement: Leverage Internet of Things (IoT) functionality to receive
data from machine owners, making the data available to their new Mobile
Apps and Web interface. Foundation for analytics to enable better insights into
the product performance, enhance service and customer experience.
➢Set up a secure service for connecting devices, ingesting machine status
using IoT Greengrass on the Edge and processing the data
➢Developed a Data Lake House platform, with transformations and a rich
API so that customers can remotely view machine data and receive
notifications of any issues with individual machines.
➢The platform is largely based on “Serverless” technologies. This was an
agile delivery, working closely with the customer product owner and front-
end team. The project took 4 months and is designed to quickly scale up to
hundreds of customers and thousands of machines
The Result:
Machine Operator
AI in Action: Connected Machines using AWS IOT

18. 1818
CASE STUDY
The Customer:
AI in Action: Demand Prediction
The Sector:
The Result:
The Solution:
The Requirement:
➢Multi ML model approach for each region/state to handle social,
economic and environmental trends that impact which model of car is
bought, i.e., hatchback vs MPV, normal drive vs 4x4, electric vs
combustion
➢An ML model per car model is then used to predict the extras that will
be ordered so that the correct configurations are built, i.e., sunroof,
colour, engine size
➢Guard rails and weighting are built to allow for business adjustments,
such as selling a new or more profitable model of car
➢Reduction in the amount of time analysts are required to look at sales
data and calculate the order for the next month. The analysts now look
at outliers and exceptions
➢Successful results in initial target regions has led to full commissioning
in every region and full productionisation
Top Global Car
Manufacturer
Automotive
Ground Stock levels are low due to supply chain
issues, therefore predicting the cars and configurations
to be built next is vital to fulfilling next quarter’s orders

19. Logical Zone architecture
Data Lake
Exploitation
Layer
Sanitised
Data sanitising rules
applied, converted to
efficient Parquet format.
Data structured and
catalogued for Data Lake
analytics use cases
Curated
Data in Normalised
and/or Data Warehouse
model. Designed for
broad consumption and
use cases. Contains
intermediate data
products
Optimised
Data Marts, Views and
optimisations applied for
specific Teams or use
cases. Modelling of this
layer aims for high
performance and contains
data products
Amazon
Athena
Landing
Data ingested in either
push or pull model into
Landing zone. Limited
structure and in same
format as source.
Amazon S3
AWS
Glue
AWS Lake
Formation
Amazon Redshift
Data
Sources
Amazon
SageMake
r
Amazon RedshiftAmazon S3
AWS
Glue
REST API
SQL
Amazon
QuickSight
AWS
Lambda
AWS IoT Core
AWS Database
Migration Service
Kinesis

20. AWS Well-Architected approach
All Analytics Solutions follow the six pillars of the AWS Well
Architected Framework. We analyse people, processes, and
technology to make informed decisions on design, technology, and
architecture. We also optimize computing services to reduce waste
and monitor performance with KPIs.
Code Efficiency
SUSTAINABILITY
Data Design &
Usage
Software
Application
Design
Platform
Deployments &
Scaling
Data Storage
Utilization
CI/CD
OPERATIONAL
EXCELLENCE
Runbooks
Playbooks
Game Days
Infrastructure
as Code
Root Cause
Analyses (RCAs)
Right AWS
Services
PERFORMANCE
EFFICIENCY
Storage
Architecture
Resource
Utilization
Caching
Latency
Requirements
Planning and
Benchmarking
Service Limits
RELIABILITY
Multi-AZ and
Regions
Scalability
Health Checks
and Monitoring
Networking
Self-healing &
Disaster Recovery
Reserved and
Spot Instances
COST
OPTIMIZATION
Volume Tuning
Service Selection
Consolidated
Billing
Savings Plans
Decommissioning
Identity and Key
Management
SECURITY
Encryption
Service Logging
and Monitoring
Dedicated
Instances and
Outposts
Compliance
Governance

21. © 2 024 Cogniza nt | Private21
Connect Data Lake Architecture

22. APPLICATIONS THAT LEVERAGE LLMs AND OTHER FMs
TOOLS TO BUILD WITH LLMs AND OTHER FMs
INFRASTRUCTURE FOR FM TRAINING AND INFERENCE
GPUs Inferentia Trainium SageMaker
EC2 Capacity Blocks NeuronUltraClusters EFA Nitro
Amazon Bedrock
Guardrails Agents Customization Capabilities
Amazon Q Amazon Q in
Amazon QuickSight
Amazon Q in
Connect
Amazon
CodeWhisperer
AWS Generative AI Stack
© 2 024 Cogniza nt22

23. © 2 024 Cogniza nt | Private23
Foundational models on AWS
AWS provides secure access to the widest range of FMs
Custom Model
Open Sources
LLMS
Use cases &
capabilities
Amazon
SageMaker
Flagship service
•A full service for
Machine Learning
•API or batch consumption
•Pay per min/hour pricing
•SageMaker has access to latest
hardware, including Inf2 and Trn1
•Cognizant has access to a wide
range of FMs (proprietary and open
source)
•Cognizant has worked with
AWS at becoming specialists
in distributed training, initially
using Hugging Face
Amazon
Bedrock
New service
•Managed service for
proprietary FMs
•Proprietary FMs require
EULA with FM author
•FMs can be fine-tuned on
your own data without sharing
your data with everyone
•Agents, Guardrails & Model
Eval
•API + Batchbased
consumption
•Now GA in limited
Regions
•Pricing per token
ProprietaryLLMS

24. © 2024 Cognizant | Private24
Gen AI seems easy but…
Source: Eduardo Ordax – Generative Lead @ AWS
Starting playing with LLM model with quick
demo/prototypes is rather “easy”
However, moving towards a Production grade
Gen AI solution is a different story…
LLM models are not NEW to us - We have
started leveraging them for 3 years
already.
And have been able to implement
production grade solution embedding LLM
models

25. Your AI and Machine Learning partner on AWS, delivering value and
innovation at the centre of your business
WHY COGNIZANT FOR GEN AI
We offer full-stack
services and a rapid,
proven path to production
and ML excellence
We help customers in a
broad range of industries
achieve their ML goals
We are trusted advisors
and practitioners of
Responsible AI
We have access to the widest
range of Foundational Models
& are experts in Fine Tuning +
Prompt Engineering
Leaders in applying Generative AI to business domains,
driving innovation and creating solutions that deliver value

26. Thank you
Philip Basford [email protected]
(He/Him)
CTO (Data & AI) – UK&I Consulting
Cognizant
AWS Ambassador -
Certification All-Star
Award
AWS Ambassador
badge image.
Experience. Advanced
level. Issued by AWS
Strategic Partner
Programs