Slides: Building Agents at Cloud Scale — Antje Barth, AWS
Source Video
Building Agents at Cloud Scale — Antje Barth, AWS
Relationship To World's Fair 2026
These slides are extracted from a public AI Engineer YouTube video connected to World's Fair 2026. Speaker-matched clips are supporting context unless later confirmed as exact session recordings; official livestream recordings are day-level/event-level source material.
Related Scheduled Sessions
- No individual scheduled session mapping has been assigned yet; treat this as an event livestream deck.
Extracted Slides
OCR text:
aWSs
ee)
@®Graphite W Windsurf) MongobB
Mdaily £3augmentcode WorkOS
OCR text:
renee er) Lg eroae
Building Agents at
Coes Cloud-Scale
Ea , /
Antje Barth (she/her)
[Wwe \ Principal Developer Advocate, Al Engineering @ AWS aws
_ ed
Ve ANTJE BARTH awWS
OCR text:
err an
ee Every customer experience
| 7 we know of
cat ,
if a Fi : :
a * will be reinvented using Al
[wo
ave ANTIJE BARTH aWS
me cee Principat Developer Advocate ~~]
OCR text:
Cc 2a
‘ amazon
ren: aes Vy , a)
. | | |
ccs ANTE BARTH aWs
OCR text:
_ oa : an
aa fi 7
= a _ yb
a |
ae he
OCR text:
a
:
uN
Urea Wit
Santa Barbara Overview ren
a
; a
a ALEXA: Santa Barbara is great for everyone.
OCR text:
: Tod
OCR text:
AIEngineer
Alexa+/Alagentsatscale
World'sFair
AEn
Honeywell
Spotify
TODAY
amazon
levoit
uber
music
BUSINES
GRUBHUB
EMERSON
INSIDES
orld'sFai
Midea
AthlooSports
TIME
Forbes
uber
Eats
alexa+
BOSE
HoseBeadtifd
hulu
fresha
ugm
Wc
GATE
dyson
VIZ10
ring
bastington
e
AP
SUNO
zoom
Govee
TIDAL
Kwikse
Jost
CONDKNANT
plex
TED
Eugust
VSVN
AEngi
WOLFRAM
XBOX
vagaro
orld's
AIEngin
Microsoft
aws
World'sFair
smol?
OCR text:
vr are Alexa+ / A! agents at scale
ae X
. 600 million devices
ny Hundreds of specialized expert systems
a ie cs p p Mi
ad Tens of thousands of partner services and devices
ane ef
Mee ei Engineering the future of Al
OCR text:
rere
oe A
i From web services
Ae to agentic services
ae [
rene as Engineering the future of Al
OCR text:
AIEngineer
World'sFair
/helpallcommands
ctrl+jnewlines
ctrl+kfuzzysearch
WhatdoestheAwSdocssayaboutAmazonBedrock?
I'llsearch the Aws documentation for informationaboutAmazonBedrockand provide you
with a summary.
AEng
orld'sFa
Using tool:search_documentation from mcp server awslabsaws_documentation_mcp_server
Runningsearch_documentation withtheparam:
arguments"
"1imit":5
"search_phrase":"Amazon Bedrock"
name":"search_documentation"
Allowthis action?Use't'to trust (always allow)this toolforthe session.
[y/n/t]:
AlEngiv
orld!
AIEngi
World'sFair
Engineering the future of Al
OCR text:
QUIZ: How long did it take to build and ship
Amazon Q Developer CLI?
me,
; aa | arama Te Ee
o 7 Se GE ae
| ‘ CRRCur area rate
OCR text:
AlEngine
World'sFair
World'sF
Launching agentic products
in months ... weeks!
AE
Worl
World'sFair
Engineering the future of Al
OCR text:
PNT teeeeeg
World's Fair
— ae ee Strands Agents
aie Sree: ——
it ran Ss gen Ss A model-driven approach to building Al agents in just a few Ines of code.
Strands Agents is an open source spine Aerts nestor seats ar vererace ee een unt vanonea'
Python SDK for building agents Mtm=wneaa<s<
iV co using just a few lines of code Feature Overview
a a fl + Leghtorenght & Fheatete ‘Sole noes sone te ue! sar vara a ta cielo
oa 6 Nate Maren eater be Amaro Bede, Ar TPeu Claman Kuntar oreders
en
aie Ll
fort.
tee pa ‘
Me eel Engineering the future of Al
OCR text:
World's Fair
wee awwew nen OOS O00. aces oem wee
cc r a - : Invoke modet !
| Promo A Get response, reasoning, tool selection
= B\Y-)
response ( Return result i
ae
Te ee Engineering the future of Al
OCR text:
ere strands EN telah
strands tools calculator
eae agent = Agent(tools=[calculator])
Wort « iir| agent( . )
oo
, ‘ F .
, { ;
|
Yoo
Ci BUILDING AGENTS AT CLOUD-SCALE
OCR text:
PN died
World's Fair strands rN eccuahe
strands.models BedrockModel
Mies ow strands.models.ollama OllamaModel
ed <a strands.models.anthropic AnthropicModel
ie strands.models.llameapi LlamaAPIModel
strands.models.openai OpenAIModel
7 strands.models.litelim LiteLLMModel
A Vy
¢ ‘fa ' ae
aie ... orimplement a custom model provider!
f . 4 |
Me ee Engineering the future of Al
OCR text:
* & File Operations - Read, write, and edit files with syntax highlighting and ntelkgent modifications
« MI Shell Integration - Execute and interact with sheti commands securely
— « ww» Memory - Store user and agent memories across agent runs to provide personalized experiences.
with both Mem0 and Amazon Bedrock Knowledge Bases
« HTTP Client - Make API requests with comprehensive authentication support
| "
; Po htay * . Slack Client - Real-time Slack events, message processing, and Slack APl access
) ig Ld “wees * 2, Python Execution - Run Python code smppets with state persistence, user conturmation for code
bee execution, and safety features
. + B Mathematical Tools - Perform advanced calculations with symbolic math capabihties
7 + AWS Integration - Seamless access to AWS services
7 \ By + F Image Processing - Generate and process images for Al applications
Ag vee * %. Video Processing - Use models and agents to generate dynamic videos
) f World 3 «© & Audio Output - Enable models to generate audio and speak
a io * te) Environment Management - Handle environment variables safely
7 x 5 « 7 Journaling - Create and manage structured logs and journals
anual « ™ Task Scheduling - Schedule and manage cron jobs
; = w Advanced Reasoning - Tools for complex thinking and reasoning capabilities
7 ‘ays » Swarm Intelligence - Coordinate muttiple Al agents for paratlel problem solving with shared
‘ hd memory
7 / + (J Multiple tools in Parallel - Cali multiple other tools at the same time in parallel with Batch Tool
rere’
Mela lek ale
OCR text:
https: //github.com/awslabs/mcp
() README 6<&% Codeotconduct ff Apache-20 tcense 87) Secunty @ = 2 Af conte butions
Deployments 500
¢ AWS MCP Servers ;
@ release
A sute of speaatzred MCP servers that help you get the most out of AWS, wherever you use MCP.
@ onnud-peges
CT 0 ee
3 rary deploy tents
Table of Contents
Languages
. AWS HCP Servers es b+
Tabie of Cortents © Lemons So, te Deckertigtin”,
Shem @ HTM.
Viihat ¢ the Model Context Protocol (MCP } and new dows # work with AV'’S MCP Seteers?
© Vihy NCP Servers?
] Avaliable Servers
= Core MCP Server
© AWS Documentaton MCP Server a
* Amazon Bedrock Knowledge Bares Retina! MCP Seever
© AWS Kendra index MCP Server hy :
© AWS COK MCP Server
© Cost Anaiyss MCP Server
# Amazon Nova Canvas MCP Server dl
OCR text:
World's Fair
coe
Sige |
ors a =X 2 . Streamable HTTP 5 Pini
enna Engineering the future of Al
OCR text:
eed Serverless MCP Server using SHTTP & AWS Lambda
World’ » °
cre
\
; \
aon ad, . Tats e ic) a IN
~ | API Gateway MCP Server
World's Fair a Microsoft 24S on?
OCR text:
eed Serverless MCP Server using SHTTP & AWS Lambda
PT ree ; Authorizer fy
7 p PN)
ef , 7 Sy a
AN : Agent , iad ’ , ic) = IN = CloudWatch
nen ee API Gateway MCP Server | ae
ft ; a - (@
caer
enna Engineering the future of Al
OCR text:
O Protocols for Agent Int yuan
a : z
ae
eee! fotos A oa a ea 5 Soe are}
AWS Glog ome = Plagne ©
AWS Open Source Blog
3 - wee bane ae. Seek. “aor wee
Open Protocols for Agent Interoperability Part 1:
Inter-Agent Communication on MCP ee Ee
tetas k AS Bye Mary Pegaiet 20d Saar: Sat bramae ait oe TYMAE SITES ete at eseaipene
Fabri grote Geureat ae A Oyen Geatee, Trught Leader Reenauca Coch | a a
SPORE e core 0)
AL AWS, ope standards run dewp in our DNA, dreving all that we do That's why we decitd ta Me Pe Es) ie Bee
bud Amazon [laste Cloud Compute ((02} as a protocok agnost« cloud computing vennce and Se oes 30 EN
Amazon SageHsker as 3 framework-agnoix deep learning service Our commitment to opteness ‘ at i : A a fs
SBE GE Te OE MOH ROSIER iba CSPRRGEST SIT ee rey da}
Protocols Rave emerged that enable the copabelity, wcluding the Vowel Content Protea! MEP ae Ts vy. + Abe fy
open sourced by Anthropic in 2024. and the Agee Tagore peotetct TAA: introduced by Google A en ele «Ste, =n)
thes year We bevieve that both MCP and AZA Rave 3 bot to offer Oemonstiating our commutment Recent Zz Toa a he rt
to theve open standards, AWS 1 josnmng Ihe steering commutes for MCP By waporting muttiole he etme Hd «br, a) oo =, os a }
theese ‘eet Er:
. as eI azat = F
~ ee xa AY Pe
; ee ANH Oe
. cate a
i a° eee ST:
[eee ren eee,
OCR text:
rere Fair '
Agent Store
—— Nee —— \ My Agent lo). —™
st ® ae
ra is : " | -
rene as Engineering the future of Al
Slide-Derived Subjects To Review
Subject extraction uses video title, related session titles/descriptions, transcript context, and OCR text when available. OCR is best-effort and should be reviewed against the embedded slide images.