Slides: Building Agents with Amazon Nova Act and MCP - Du'An Lightfoot, Amazon (Full Workshop)
Source Video
Building Agents with Amazon Nova Act and MCP - Du'An Lightfoot, Amazon (Full Workshop)
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:
=, ae
AMtngineer
World's Fai
OCR text:
Building Agents
with Amazon Nova
iw Act and MCP
Du’An Lightfoot Banjo Obayomi
Sam OCT ote Tol eC Tamera ATMS EME
aws
ed
OCR text:
rs aie
What is Agentic Al?
Memory fe] estar stele) |
Tools _
Actions Environment
. Agent Brain
a Al agents are autonomous systems that understand,
aoe plan, and execute actions to achieve specific goals.
Oals
OCR text:
What is Agentic Al?
ttalela | loleet areola] |
Afetolks a“ :
Actions Environment
. Agent Brain
a Al agents are autonomous systems that understand,
oan plan, and execute actions to achieve specific goals.
Vr cai)
| a Microsoft § = oct)
OCR text:
What does an agentic system look like?
Agent workflow 1)
Agent components | ea 2 )
, Fs
; a Microsoft Qa?
OCR text:
a a A ac an a a In i
What does an agentic system look like?
ran acd diel od 1)
Agent components | ie 2)
a f+ ° —
eae Peer ea Sera Canaan PereEreets ore arrerraan 2m
aera are
eee AMM Le ER eS : 3)
rar
Ree Oe ee eee ees
hid
aWS
ee)
OCR text:
ne ain
‘ - ?
What does an agentic system look like:
Agent workflow 1)
7 an
oon ere ane
Sytann
Vrs eaodta dearaar es | ee 2)
a a : = . , eee
Pern : —_ 7 : 7 : ao a
eae Perera eae Crean erent ren coer are tary eT bad -
i) . ’ CO ae
RE nr eared eee ea prea
Cael eee Aaa se etn ced : 3) oo |
n elt) eee aaa Corer a
: ner:
rer
Seen Grant] apnea crea’
Rr eT tt See te Sea ELS ial ad Coen ce
a
re : 7 2
| eee)
OCR text:
es a eet
Use agentic Al for: Traditional solutions are
Better for:
* Complex workflows with variable paths » Fired cule basec workfiews wth
SS pred ctable path to ovtcemes
« Coordination across multiple tools and : : MaNUCER CER oiementiricnr
Se Ries fealaastradenelgeteaeta)
« Tasks requiring understanding context MEE SE a cleliet ciafeN PESd stone hidelan hs 1atela eect
and adapting responses a ; ;
; Camere eee abana: Way aCe ae Ma re Cea nT
+ Long-running processes needing memory oO Tet reeen ne
across sessions
AAP SS
entities
“eed
OCR text:
a saan “ ae
SPECIALIZED FULLY-MANAGED o1hd
rare n el Amazon Bedrock Agents STO h et clei sd
Pre-built products that use with bult-in FM- powered Simple, flexible, and
agents for enhanced orchestration lightweight open source
productivity SOK for building agents
OUT-OF-THE-BOX AGENTS TOOLS FOR BUILDING
Pa as
a : - A
Pe)
i Dekel ro) aS 008 OO
OCR text:
TT Ease of Use: Intuitive agent
development to get started in
minutes instead of hours
te Robust Capabilities: Native tools,
Stra nd S Age ales oe MCP servers, and natively integrated
PE Nese adl at)
Strands Agents is an open-source
python SDK for building agents
using just a few lines of code nex Extensibility: Support for custom model
Qo providers, custom tools, and MCP
</> Rapid Development: Quick
prototyping and iteration
OCR text:
Strands Agents
AgenticLoop
AIE
Invoke model
Prompt
Invoke agent
Model
Response,reasoning,
Agent
toolselection
Final response
Execute tool
Result
Tools
Result
aws
OCR text:
Nema
eee ne
Ce mee nape een >
A nemper wasn @ sa Toney Ae
7 os ww eee. = wee we ae
8 for Mant i Raderecd Cay, CA gle @ wm
jadd te fw © ey
| ~w f Mo Seas
eee eee Mee a.
a a --
SS WAR aye. a eR Se rm wee eee .
i? re : ee Deena ms Z
i
F fai)
Sicko ES 08 O8
OCR text:
What is Model Context Protocol (MCP)
Client a LS
Knowledge base
, GitHub
a Microsoft ary?
OCR text:
Let’s build!
https://s12d.com/ai_eng_aws
OCR text:
aws
aws
World'sFair
OCR text:
Let's build!
ae https://s12d.com/ai_eng_aws
| fea] aws
7
OCR text:
Sh er i i i id a sie sii
ee Gee ee * @S teens mE + 19 capes ONES: GEES a eI ey ge eee _ 3
. . a eran ves i : ry
Buliding Agents ath «
fesaton' Nove Ait snd We sccene ta the Thakteng Ager arth Amaron feos Act and HOP wort ney 6 ths hands on weston, your inaen hie ho Dusit encatie atom tat ng agents
wee
WOU eIpiere how 10 Cte IRAE IRE Can Rav ate wtb ent aCER. perform COmplES ALE. BE emegy ate wth Wesel Conrtent Protos IMAP wervery Bee enhance Capable es
cree tony wR :
D mreates tere a
Men eo Reo nb art ogee
Sara o
Heke: Race wt Mate
Meee 1 marin at eae
ae ‘
wrens oe
vere
sree oe LE
pareveg tnt
perees re t
sae are tt
ri bgt ne
7 i oN F Zi — -
To os Dea hoe S12) Ce |
ey
7 rey |
OCR text:
eo 6 (Oe Eee 6 gees: 8 aesrerg oy a: OPE eee a eee . °
. . : morn ces pes & : Oe
Agents eth Amarone eee rer - eo
Meva Act and MCP ad
Later Sacco ellen Ageres OD Amat Marea ASE and ME
beta Age woh aasor font .
Aree Al Engineer World's Fair 2025: Building Agents with Amazon Nova Act and MCP
ciate Coin yet
# mene tere Event information
woos Rute Att ano vert ene owwwes seen repent
eee BER Proy t2 ap ene stron eet wees?
Moats Micnz ert Mose
are Pere 2 A
Mette E Racony wrt arth ‘Gearwaes,
: trace Nee De ees ty enn Naren Nes tw ht ee Cane eed RP eerste sent ad ats natn tenes a Poor Noes ets moan ogee
= tac setties b pniry Mee SA8 forme AN NO Mal NO Nd AP gt Nee OF et AOL Ym Mt aD ind MOR IH EAE RA PL Va Dee otra CORD wt EME KCC es
mace Bee ark Od Tema Ned ang ange re tes
wl
r .
Workshop Le atee >
wrrawes 200 | Re ittoterted JL)
ewes oe oe Te. Conmptaratty eves A ere rc
Raarg Age em Amn AT aI O Arana tect sereriteet be
mvs not
Araceae LE
a ete eal? Dewrrotion
Deco Som te Ded opie my Sembrng Arden Names At wt? he Mace: Sere Proton Gt OAL er OCHS AGRE! TRUE NAG RODRIG ne perso Lormplen KAUN Tron Siok Act Brean asoraen
(apes aptent Men AS Haren AapetN Oa Te OME A TS ans fe one Nek te INGE Owe Menem te Ind Rp>HE Or me Ben ori dene eal tate Hee oat me
sent ee ol ena ald aenpege mer nsers
© AWS acconet acces
enact fe
Event Outputs 7) ©)
ter AT C4 Caer
wand Annes at ts ’ *
vat © emai 2 og < om ©
Ee a F mt "
LT) ‘a: a7r00 Ce |
OCR text:
ial en a eS eee bs —_
ee RG REE 8 ee: @ BD eeesee coy 0 BD German oe qpevere se @ ONS @ OME Soe GW. tae ee #
. no were: : : : oe
@ mv rens e
2 Neeteery saat oo
mw Oe eee Teo i ran
aaa, Pasa ate aN CY eee cee Cee eaten eaten en ere eter iied ea
ne! a :
© teundsoee models
seceecregy mew Overview -..
Mette States ew
Maing reed fe eg Model wpotiegnt
rer ee A\ Arahropic’s Claude
Payprennts ere Peet set cee eared far Pentre |
We ee a et Pe A he ald ae
ste “Er OE et eet are pele AOR
nen ten woes
(© Bulhdas tools poet mee mers
sen
, R
tareerge tate é fo
rigs taneser
© Sotequacds
aeres Chat / Tent tenage / Video
SER SAS at eat a Mae He 8 and LANGE LPO GE LAR TA nares foe Freed ROCA Thy Can ene 5 Ung Haney (rterite carga Phage Oy DNS Tae AMOR TE few Canine Mader I The Baypred, sate 2 ter
© tnterence ond Anervimmre Drone oF or the meet fy DING vahargvert DeTOm Hhel Cane mie ORC The (OTE HF Brena protest oo wr torte
TE a perenne rncornen © 8 Coat Sarwat
Fre eteel Ste!
peeenee pee Mergen Meet erent
won taper tere
Builder tools
ite enh tO Ee ete Vy Wn ne eee on tek el ener emis att ere
. aN . 5 — -
LT as ee oe SS) Cr ae |
OCR text:
z . me weenie oon, : & : en
e=. oc. oz
Workshop Search Results
Lees decane etn unin va (© UOFh, deneeen Wet Sorte, Ot & Os alain AB igh eervel. Privacy | $60 Somes | Concert tet
: je" °8 9°00 Cn aan |
a -
a Microsoft Gnu
OCR text:
ee ee ai ee —
Se oes gebo “es
EE Tela]
AE a] oe
Ore 4
ane
a bee TPE
7 Lace Mae oOCT abet Sd lacramost liv can ieee some)
ar) a
Se” °*8 0700 GQ =
mV ai)
Microsoft §=OQ7aoU
OCR text:
SIU AYLSN7
(a) Fie Del
ee
td oe
2 os al
2: ie T-Te
ial sens
https://pulse.aws/promotion/RPND9AFU?p=0
a Microsoft aur?
OCR text:
P|
Survey sea
dsl
or
im rd AO:
r
https://pulse.aws/promotion/RPNDSAFU?p-0
, aoe a _ ; ee
po i Onn Dal i<lo ct Bs
a Microsoft ary?
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.