Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
フリマアプリ「メルカリ」AIチームの サービスライフサイクルと、 メルカリにおけるフルサイクル...
Search
Shopetan
September 10, 2020
Technology
0
1.2k
フリマアプリ「メルカリ」AIチームの サービスライフサイクルと、 メルカリにおけるフルサイクルエンジニア / bitvalley2020-devdays-shopetan
BIT VALLEY 2020 9/10(木) Developers Days
Track 2 14:55-15:15 @ss_shopetan
Shopetan
September 10, 2020
Tweet
Share
Other Decks in Technology
See All in Technology
非情報系研究者へ送る Transformer入門
rishiyama
13
8.4k
Claude Code のコード品質がばらつくので AI に品質保証させる仕組みを作った話 / A story about building a mechanism to have AI ensure quality, because the code quality from Claude Code was inconsistent
nrslib
13
8.6k
ReactのdangerouslySetInnerHTMLは“dangerously”だから危険 / Security.any #09 卒業したいセキュリティLT
flatt_security
0
310
Tebiki Engineering Team Deck
tebiki
0
27k
JAWSDAYS2026_A-6_現場SEが語る 回せるセキュリティ運用~設計で可視化、AIで加速する「楽に回る」運用設計のコツ~
shoki_hata
0
3k
Agent ServerはWeb Serverではない。ADKで考えるAgentOps
akiratameto
0
110
銀行の内製開発にて2つのプロダクトを1つのチームでスクラムしてみてる話
koba1210
1
140
Oracle Cloud Infrastructure IaaS 新機能アップデート 2025/12 - 2026/2
oracle4engineer
PRO
0
170
TypeScript 7.0の現在地と備え方
uhyo
7
1.7k
ソフトバンク流!プラットフォームエンジニアリング実現へのアプローチ
sbtechnight
1
180
AIエージェント、 社内展開の前に知っておきたいこと
oracle4engineer
PRO
2
150
決済サービスを支えるElastic Cloud - Elastic Cloudの導入と推進、決済サービスのObservability
suzukij
2
660
Featured
See All Featured
DBのスキルで生き残る技術 - AI時代におけるテーブル設計の勘所
soudai
PRO
64
51k
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
47
8k
Breaking role norms: Why Content Design is so much more than writing copy - Taylor Woolridge
uxyall
0
210
16th Malabo Montpellier Forum Presentation
akademiya2063
PRO
0
73
Collaborative Software Design: How to facilitate domain modelling decisions
baasie
0
160
The World Runs on Bad Software
bkeepers
PRO
72
12k
A Modern Web Designer's Workflow
chriscoyier
698
190k
How to Get Subject Matter Experts Bought In and Actively Contributing to SEO & PR Initiatives.
livdayseo
0
85
Optimising Largest Contentful Paint
csswizardry
37
3.6k
AI in Enterprises - Java and Open Source to the Rescue
ivargrimstad
0
1.2k
Understanding Cognitive Biases in Performance Measurement
bluesmoon
32
2.8k
Rails Girls Zürich Keynote
gr2m
96
14k
Transcript
1 ϑϦϚΞϓϦʮϝϧΧϦʯAIνʔϜͷ αʔϏεϥΠϑαΠΫϧͱɺ ϝϧΧϦʹ͓͚ΔϑϧαΠΫϧΤϯδχΞ BIT VALLEY 2020 9/10() Developers Days
Track 2 14:55-15:15 ٠ ฏ (Twitter: @ss_shopetan)
2 ࣗݾհ !TIPQFUBO !TT@TIPQFUBO .FSDBSJ1FSTPOBMJ[BUJPO$BNQ3FDPNNFOEBUJPO5FBN &YQFSJFODF *OUFSOTIJQ.-&OHJOFFS
.FSDBSJ 1SFTFOU4PGUXBSF&OHJOFFS .FSDBSJ -JLF &TQPSUT8PSLPVU,BSBPLF$PODFSU
3 ͜ͷϓϨθϯͰͷΰʔϧ Non-Goals Goals ಛఆͷٕज़తͳղܾͱͦ ͷΞϓϩʔνΛΔ ಇ͖ํʹؔ͢Δ ྑ͠ѱ͠Λٞ͢Δ͜ͱ ਪનυϝΠϯΛ௨ͨ͡ ͓٬༷ͷαʔϏε
ϥΠϑαΠΫϧΛΔ ϑϧαΠΫϧΤϯδχΞʹ ͍ͭͯΔ
4 ̏ͭͷࣄۀల։ 4
5 5 ςΫϊϩδʔͷྗͰՁަͷ͋ΓํΛม͑ΔαʔϏε ͜Ε·Ͱ େྔੜ࢈ɾେྔফඅ ୭͔ʹͱͬͯඞཁͳͷͰ ࣺͯΒΕͯ͠·͏ ඞཁͳͷ͕ඞཁͳਓʹ ඞཁͳ͚ͩಧ͘ ͜Ε͔Β
ͷ ! ςΫϊϩδʔʹΑͬͯͰ͖Δ҆৺ɾ҆શ ◦ ΤεΫϩʔܾࡁ ◦ AIٕज़Λۦͨ͠ҧݕγεςϜ ! ʮചΔ͜ͱΛۭؾʹ͢ΔʯUXͱςΫϊϩδʔ ◦ όʔίʔυग़ ◦ AIը૾ೝࣝ ϝϧΧϦͱ͍͏CtoCϚʔέοτϓϨΠε
6 Values େʹΖ͏ શͯޭͷͨΊʹ ϓϩϑΣογϣφϧͰ͋Ε Go Bold All for One
Be a Pro
7 ϝϧΧϦͷAIٕज़ʹ͍ͭͯ ϝϧΧϦͰͷAIٕज़ ԿʹΘΕ͍ͯΔ?
8 ϝϧΧϦͷAIٕज़ʹ͍ͭͯ(AIग़ / ը૾ݕࡧ) ײಈग़ͷىҊ͔ΒϦϦʔε·Ͱ
9 ϝϧΧϦͷAIٕज़ʹ͍ͭͯ(AIग़ / ը૾ݕࡧ) ϑϦϚΞϓϦʮϝϧΧϦʯɺʮࣸਅݕࡧػೳʯΛಋೖ
10 ϝϧΧϦͷAIͰղܾ͖ͯͨ͠ྖҬʮग़ʯ Πϯετʔϧ/ ձһొ ͢Δ ചΓͨ͘ͳΔ ௐࠪ͢Δ [͚ͳͲ] ങ͍ͨ͘ͳΔ ݕࡧ/ൺֱ͢Δ
ग़࡞ۀ͢Δ ަব/ߪೖ͢Δ Attention Search/ Compare Interest Action ૹ࡞ۀ͢Δ डऔධՁ͢Δ Ship/ Evaluate ɾAIग़ ɾΧλϩάϚοϐϯά ɾՁ֨αδΣετ ɾݕࡧΞϧΰϦζϜ ɾಉٛޠ֦ு ɾࣸਅݕࡧػೳ ɾҧग़ͷࢹ ɾҟৗߦಈͷݕ ɾૹαδΣετ ɾॏྔਪఆ
11 ʮ͓٬༷ʯͷతͳମݧ্ʮʯҎ্ʹॏཁʹͳΔ ݟ͔ͭΔ ݟ͚ͭΔ ϢϏΩλε ΞϯϏΤϯτ ϝϧΧϦ$10͕ߟ͑Δɺ"*࣌ͷϏδωεͷม༰
12 ͓٬͞·ͷʮಈػ͚ͮʯΛࢧԉ͢Δ Πϯετʔϧ/ ձһొ ͢Δ ചΓͨ͘ͳΔ ௐࠪ͢Δ [͚ͳͲ] ങ͍ͨ͘ͳΔ ݕࡧ/ൺֱ͢Δ
ग़࡞ۀ͢Δ ަব/ߪೖ͢Δ Attention Search/ Compare Interest Action ૹ࡞ۀ͢Δ डऔධՁ͢Δ Ship/ Evaluate τϨϯυ / ϥϯΩϯά ͷਪન ͓٬༷ͷߦಈʹԠͨ͡ ਪન ϨίϝϯυΛ௨ͯ͡ ʮಈػ͚ʯΛڧԽ͢Δ
13 Ϩίϝϯυ͕ͨΒ͢٬؍తͳΠϯύΫτ Alibaba͕ӡӦ͢ΔC2CαʔϏεຊ ԁͰ2400ԯԁنͷऩӹ ͦͷ͏ͪϨίϝϯσʔγϣϯػೳ͕ऩӹ ͱτϥϑΟοΫͷେ෦ΛΊ͍ͯΔ ϝϧΧϦͷϨίϝϯσʔγϣϯػೳશ ମτϥϑΟοΫͷ͏ͪ10-20%ڧɹɹ Big ImpactΛͨΒ͢ྖҬ
#JMMJPOTDBMF$PNNPEJUZ&NCFEEJOHGPS&DPNNFSDF3FDPNNFOEBUJPOJO"MJCBCB
14 ϨίϝϯσʔγϣϯʹऔΓΉ·Ͱͷ γεςϜత ৫త ϞϊϦεͳ ΞϓϦέʔγϣϯͰӡ༻ վળίετ͕ߴ͍ Β͘ϝϯςφϯε ͞Ε͍ͯͳ͍ ϨίϝϯυΛઐͱ͠
Productૌٻ͢Δ৫ମ੍ ͕ແ͔ͬͨ AIνʔϜʹ ڠۀମ੍͕͔ͬͨ͠
15 ղܾํ๏?
16 ղܾํ๏? Ϩίϝϯσʔγϣϯʹؔ͢Δ͋ΒΏΔ ݖݶͱΛ࣋ͬͨνʔϜΛ࡞Δ ➕
17 ղܾํ๏? Ϩίϝϯσʔγϣϯʹؔ͢Δ͋ΒΏΔ ݖݶͱΛ࣋ͬͨνʔϜΛ࡞Δ ͋ΒΏΔαʔϏεϥΠϑαΠΫϧ͕ νʔϜͰ݁͢ΔΈΛ࡞Δ ➕
18 ղܾํ๏? Ϩίϝϯσʔγϣϯʹؔ͢Δ͋ΒΏΔ ݖݶͱΛ࣋ͬͨνʔϜΛ࡞Δ ͋ΒΏΔαʔϏεϥΠϑαΠΫϧ͕ νʔϜͰ݁͢ΔΈΛ࡞Δ ➕ MicroserviceԽ
19 ։ൃνʔϜͷMicroserviceԽͱ NJDSPTFSWJDFTQMBUGPSNBUNFSDBSJ Develop Monolith Test Deploy Operate Backend Team
QA team SRE team
20 ։ൃνʔϜͷMicroserviceԽͱ NJDSPTFSWJDFTQMBUGPSNBUNFSDBSJ Test Deploy Operate Backend Team QA team
SRE team Develop Service B Develop Service A Develop Service C
21 ։ൃνʔϜͷMicroserviceԽͱ NJDSPTFSWJDFTQMBUGPSNBUNFSDBSJ Test Deploy Operate Backend Team QA team
SRE team Develop Service B Develop Service A Develop Service C
22 ։ൃνʔϜͷMicroserviceԽͱ NJDSPTFSWJDFTQMBUGPSNBUNFSDBSJ Test Deploy Operate Backend Team QA team
SRE team Develop Service B Develop Service A Develop Service C
23 ։ൃνʔϜͷMicroserviceԽͱ NJDSPTFSWJDFTQMBUGPSNBUNFSDBSJ Develop Service B Test Deploy Operate Service
B Team Develop Service A Test Deploy Operate Service A Team Develop Service C Test Deploy Operate Service C Team
24 ։ൃνʔϜͷMicroserviceԽͱ NJDSPTFSWJDFTQMBUGPSNBUNFSDBSJ Develop Service B Test Deploy Operate Service
B Team Develop Service A Test Deploy Operate Service A Team Develop Service C Test Deploy Operate Service C Team ։ൃ͔Βӡ༻·ͰϑϧαΠΫϧʹಇ͘
25 ϑϧαΠΫϧʹಇ͘ͱ https://medium.com/netflix-techblog/full-cycle-developers-at-netflix-a08c31f83249
26 ϝϧΧϦͷ߹ MicroserviceԽΛ͢Δ ≒ ϑϧαΠΫϧΤϯδχΞͱͯ͠ৼΔ͏
27 AIνʔϜͰͷαʔϏεϦϦʔε·ͰͷϥΠϑαΠΫϧ Test Deploy Operate ML Engineer MLOps team Develop
Model B Develop Model A Develop Model C
28 AIνʔϜ֤υϝΠϯ͝ͱʹνʔϜׂͨ͠ Develop Recommend Platform Test Deploy Operate Recommendation Team
Develop Image search Test Deploy Operate Image search Team ML Engineer MLOps Team Search Team
29 ϨίϝϯσʔγϣϯνʔϜͷαʔϏεϥΠϑαΠΫϧ Develop Recommend Platform Test Deploy Operate Recommendation Team
30 ϨίϝϯσʔγϣϯνʔϜͷαʔϏεϥΠϑαΠΫϧ Develop Recommend Platform Test Deploy Operate Recommendation Team
ͲΜͳׂ͕ඞཁ?
31 ϨίϝϯσʔγϣϯνʔϜͷαʔϏεϥΠϑαΠΫϧ Develop Recommend Platform Test Deploy Operate Recommendation Team
Act as SWE, SET, SRE
32 ϨίϝϯσʔγϣϯνʔϜͷαʔϏεϥΠϑαΠΫϧ Develop Recommend Platform Test Deploy Operate Recommendation Team
MLϓϩδΣΫτͳΒ?
33 ϨίϝϯσʔγϣϯνʔϜͷαʔϏεϥΠϑαΠΫϧ Develop Recommend Platform Test Deploy Operate Recommendation Team
Act as ML, BI, PM Survey Analyze Business Impact Modeling Experiment Design Develop Platform
34 ϨίϝϯσʔγϣϯνʔϜͷαʔϏεϥΠϑαΠΫϧ Develop Recommend Platform Test Deploy Operate Recommendation Team
Survey Analyze Business Impact Modeling Experiment Design Develop Platform ։ൃνʔϜ͕ SWE/SET/SRE/ML/BI/PM ΛϑϧαΠΫϧʹ୲͢Δ
35 ϨίϝϯσʔγϣϯυϝΠϯͷMicroserviceԽͷ֓ཁ
36 MicroserviceҠߦͯ͠ಘΒΕͨ͜ͱ γεςϜత ৫త Scalability Availability ‐‐‐ Complexity վળ Velocity,
Agility‐
37 MicroserviceҠߦͯ͠ಘΒΕͨ͜ͱ γεςϜత ৫త Scalability Availability ‐‐‐ Complexity վળ Velocity,
Agility‐ ਅʹ࣮ݱ͍ͨ͜͠ͱ ʮΤϯδχΞϦϯάΛ௨͓ͨ͡٬༷ͷܧଓతͳՁఏڙʯ
38 [࠶ܝ] MicroserviceԽͰऔΓΉ͜ͱࣗମՄೳʹ Ϩίϝϯσʔγϣϯʹؔ͢Δ͋ΒΏΔ ݖݶͱΛ࣋ͬͨνʔϜΛ࡞Δ ͋ΒΏΔαʔϏεϥΠϑαΠΫϧ͕ νʔϜͰ݁͢ΔΈΛ࡞Δ ➕ MicroserviceԽ
39 [࠶ܝ] ϝϧΧϦͷ߹ MicroserviceԽΛ͢Δ ≒ ϑϧαΠΫϧΤϯδχΞͱͯ͠ৼΔ͏
40 ϝϧΧϦͰMicroserviceԽ͢Δਅͷར
41 ϝϧΧϦͰMicroserviceԽ͢Δਅͷར ϑϧαΠΫϧΤϯδχΞͱͯ͠ৼΔ͑Δ ‑
42 ϝϧΧϦͰMicroserviceԽ͢Δਅͷར ϑϧαΠΫϧΤϯδχΞͱͯ͠ৼΔ͑Δ ‑ νʔϜ͕αʔϏεϥΠϑαΠΫϧʹ ΦʔφʔγοϓΛ࣋ͯΔ
43 αʔϏεϥΠϑαΠΫϧʹΦʔφʔγοϓΛ࣋ͯΔͱ Develop Recommend Platform Test Deploy Operate Recommendation Team
Microservice Migrate Done
44 αʔϏεϥΠϑαΠΫϧʹΦʔφʔγοϓΛ࣋ͯΔͱ Develop Recommend Platform Test Deploy Operate Recommendation Team
Velocity, Agility ⬆
45 αʔϏεϥΠϑαΠΫϧʹΦʔφʔγοϓΛ࣋ͯΔͱ Develop Recommend Platform Test Deploy Operate Recommendation Team
Scalability, Availability ⬆
46 αʔϏεϥΠϑαΠΫϧʹΦʔφʔγοϓΛ࣋ͯΔͱ Develop Recommend Platform Test Deploy Operate Recommendation Team
Scalability, Availability ⬆ ͜Εࣗମຊ࣭Ͱͳ͍
47 αʔϏεϥΠϑαΠΫϧʹΦʔφʔγοϓΛ࣋ͯΔͱ Develop Recommend Platform Test Deploy Operate Recommendation Team
Survey Analyze Business Impact Modeling Experiment Design Develop Platform
48 αʔϏεϥΠϑαΠΫϧʹΦʔφʔγοϓΛ࣋ͯΔͱ Develop Recommend Platform Test Deploy Operate Recommendation Team
Survey Analyze Business Impact Modeling Experiment Design Develop Platform
49 αʔϏεϥΠϑαΠΫϧʹΦʔφʔγοϓΛ࣋ͯΔͱ Develop Recommend Platform Test Deploy Operate Recommendation Team
Survey Analyze Business Impact Modeling Experiment Design Develop Platform νʔϜ͕End-to-EndͰResponsibilityΛ࣋ͯΔ
50 αʔϏεϥΠϑαΠΫϧʹΦʔφʔγοϓΛ࣋ͯΔͱ Develop Recommend Platform Test Deploy Operate Recommendation Team
Survey Analyze Business Impact Modeling Experiment Design Develop Platform E2EͰResponsibilityΛ࣋ͯΔࣄͰ ϥΠϑαΠΫϧશମͷVelocity্͕͕Δ
51 αʔϏεϥΠϑαΠΫϧʹΦʔφʔγοϓΛ࣋ͯΔͱ Develop Recommend Platform Test Deploy Operate Recommendation Team
Survey Analyze Business Impact Modeling Experiment Design Develop Platform ʮΤϯδχΞϦϯάΛ௨͓ͨ͡٬༷ͷܧଓతͳՁఏڙʯ Λߴʹ࣮ݱ͠ଓ͚Δ͜ͱ͕ՄೳʹͳΔ
52 ͱ͍͑… Develop Recommend Platform Test Deploy Operate Recommendation Team
Survey Analyze Business Impact Modeling Experiment Design Develop Platform
53 ͱ͍͑… Develop Recommend Platform Test Deploy Operate Recommendation Team
Survey Analyze Business Impact Modeling Experiment Design Develop Platform ։ൃνʔϜ͕ SWE/SET/SRE/ML/BI/PM ΛϑϧαΠΫϧʹ୲͢Δ 4VQFS)BSEͰʜ
54 ͳͥ͜Ε͕࣮ݱͰ͖Δͷ͔ʁ ϝϧΧϦ$50໊ଜ͕ࢦ͢ʮ౷ͷͱΕͨ༗ػతͳ৫ʯͱʁ
55 ϨίϝϯσʔγϣϯνʔϜͱͯ࣌͠ͷTL͕ܾΊͨ͜ͱ ! AI͔Βͱ͍͏ઢΛΊΔ ◦ MLΤϯδχΞͰ͋ΔҎલʹιϑτΣΞΤϯδχΞͰ͋Δ ◦ ͋͘·ͰखஈͳͷͰඞཁͳλΠϛϯάͰར༻͢Δ ◦ අ༻ରޮՌɼ൚༻ੑΛҙࣝ͠ΠϯύΫτͷ͋Δ෦͔ΒMVPʹ࣮ݱ
◦ Decide Everything Scientifically ! طଘͷׂͷڥքΛͳͨ͘͠νʔϜΛ࡞Δ ◦ ݩͷॴଐҧ͑Ͳతಉ͡ ◦ ڥքෆ৹ΛੜΉ ◦ ઓ͍ͨ͠ਓ͕ઓ͍ͨ͠ྖҬʹऔΓΊΔνʔϜʹ͢Δ ◦ ಘҙྖҬͱνϟϨϯδͷόϥϯεΛอͪͳ͕Β։ൃͨ͠
56 ͱ͍͑ࣗʹೳྗ͕͋ͬͨͷ͔? ! ݩʑMLΤϯδχΞͱͯ͠Πϯλʔϯ -> ೖࣾ ◦ SWEɼSETɼSREͱͯ͠ͷҰൠతͳڭཆ͢Βͳ͔ͬͨ ◦ MicroserviceͷMigrationવ͕ͩೖࣾޙॳΊͯܞΘͬͨ
! ϝϧΧϦͷจԽͱνʔϜͷํ͕ࣗΛٹͬͯ͘Εͨ ◦ GoBoldʹະܦݧͷྖҬʹઓͯ͠ྑ͍ڥͩͬͨ ◦ ͓٬༷ͷՁఏڙΛୈҰʹߟ͑ͯɼ Ұ࿈ͷαʔϏεϥΠϑαΠΫϧΛҰਓͰ୲Ͱ͖ΔΑ͏ʹͳͬͨ
57 ΦʔφʔγοϓΛ࣋ͯͯΑ͔ͬͨ͜ͱ(Team) ! Ϩίϝϯσʔγϣϯʹؔ͢ΔશͯͷυϝΠϯΛվળ͢Δͱݖݶ͕͋Δ ◦ طଘͷϨίϝϯσʔγϣϯʹؔ࿈͢ΔશͯͷAPI͕ ϨίϝϯσʔγϣϯνʔϜͷϓϥοτϑΥʔϜͱܨ͕͍ͬͯΔ ◦ ܧଓతʹ࣮ݧͱվળΛ܁ΓฦͤΔڥ !
ఏى͔ΒϦϦʔε·ͰͷվળαΠΫϧʹ֎෦ґଘ͕গͳ͍ ◦ ϨίϝϯσʔγϣϯνʔϜͰطʹӡ༻͍ͯ͠ΔAPIܭը͔Β ϦϦʔεஅ·ͰνʔϜ෦ͷΈͰ݁͢Δ ◦ ৽نͷAPIଟͯ͘1-2νʔϜͱͷௐ͚ͩͰຊ൪ϦϦʔε͕Ͱ͖Δ
58 ΦʔφʔγοϓΛ࣋ͯͯΑ͔ͬͨ͜ͱ(ݸਓ) ! MicroserviceԽ/Platform։ൃʹΑΔSWEͱͯ͠ͷٕज़ྖҬͷ֦େ ◦ Manage Elasticsearch on GKE ◦
Develop middleware ! PlatformΛӡ༻͢Δ͜ͱʹΑΔSREͱͯ͠ͷҙࣝ ◦ Monitor any metrics ◦ Consider SLI/SLO for Recommendation ◦ Loadtest with Real-Scenario ◦ Consider Cluster’s constitution ◦ Apply CI/CD ◦ Infrastructure as Code ◦ υϝΠϯશମͰͷCostཧ
59 ϑϧαΠΫϧʹಇ͚ΔΑ͏ʹͳ͕ͬͨ… ! Q. ࣮ࡍʹϑϧαΠΫϧͳϑϩʔͰಇ͍͍ͯΔ? ◦ A. Yes! ◦ ϦϦʔεӡ༻·ͰҰ؏ͯ͠୲͢Δ͜ͱ͕Մೳ
! Q. νʔϜͰϑϧαΠΫϧҎ֎ͷಇ͖ํ͋Δ? ◦ A. Yes! ◦ ࣗࣗɼڵຯͷ͋ΔྖҬ͕αΠΫϧࣗମͷվળʹ͋Δ ▪ Πϯελϯεͷࣗಈఀࢭ/GKE NodeࣗಈఀࢭͳͲͷίετΧοτ ▪ K8s্Ͱͷscalable, ࠶ݱੑͷߴ͍Loadtest clientͷ࣮ݱ ▪ ElasticsearchΛ͡Ίͱͨ͠ϓϥοτϑΥʔϜͷࢹػߏͷՃ ▪ ͜ΕΒͷIaCରԠͳͲ
60 αʔϏεϥΠϑαΠΫϧΛࢧԉ͢Δ ྫ͑ : AI Platform(πʔϧ) - ModelཧPipelineΛࢧԉ - ABςετࢧԉ
ྫ͑: Embedded SRE(ׂ) - υϝΠϯʹಛԽͨ͠SRE ϥΠϑαΠΫϧʹؔΘΔ༷ʑͳ ΛղܾΛ͢Δ͜ͱ͕ࣄ
61 ఆ͑͠Δ࣭ ! MLͷSpecialistͱͯ͠ࣄͯͨ͠ͷʹશͯͷϑΣʔζΛΒͳ͍ͱμϝ? ◦ No. ͦΜͳ͜ͱͳ͍ ◦ ֤ʑʹಘҙͳྖҬ͕͋ΔͷવͳͷͰํνʔϜ͝ͱʹҙࢥܾఆ ◦
νʔϜશମͰe2eʹResponsibilityΛ࣋ͯΔ͜ͱʹՁ͕͋Δ ! ClientFrontendͳͲͷྖҬ͕ඞཁͳ࣌Ͳ͏͢Δͷ? ◦ ϝϦΧϦͰCampͱ͍͏୯ҐͰ։ൃΛ͍ͯ͠Δ ▪ ؔ࿈͢ΔྖҬʹΛ࣋ͭνʔϜ͕ू߹͍ͯ͠Δ ! Include all tech stack ▪ Q͝ͱͷඪΛCampͱ͍͏୯ҐͰڞ༗&CampͰεΫϥϜ ! ༏ઌΛἧ͑ͯ։ൃ͢ΔͷͰϦϦʔεεϐʔυ͕อূ ͓ؾ͖ͮͰ͔͢ʁύʔιφϥΠθʔγϣϯ͔Βੜ·ΕͨϝϧΧϦػೳվળ·ͱΊ
62 ·ͱΊ ! ਪનυϝΠϯͷαʔϏεϥΠϑαΠΫϧΛհ ◦ جຊతʹMLʹؔ͢ΔProductͷ্ཱͪ͛ͱมΘΒͳ͍ ◦ MicroserviceԽͷརe2eʹΦʔφʔγοϓΛ࣋ͯΔࣄ ▪ νʔϜͰఏى͔Βҙࢥܾఆ·Ͱ݁͢Δ
▪ AIྖҬʹ͓͍ͯҰ࿈ͷTry&ErrorͷߴԽʹՁ͕͋Δ ! ϑϧαΠΫϧΤϯδχΞͱͯ͠ಇ͍͍ͯΔνʔϜΛհ ◦ ֤ΤϯδχΞ͕શͯͷྖҬΛۃΊͳ͍ͱ͍͚ͳ͍༁Ͱͳ͍ ◦ GoBoldʹઓ͍ͨ͠ਓ͕ઓ͍ͨ͠ྖҬʹऔΓΊΔΑ͏ʹܾΊͨ ◦ αΠΫϧࣗମΛվળ͢Δ͜ͱʹՁ͕͋Δ ◦ ϑϧαΠΫϧʹϝϧΧϦͷϨίϝϯυΛΑ͍͖ͯ͘͠·͢
63 #BDLFOE J04 "OESPJE 'SPOUFOE 4JUF3FMJBCJMJUZ .BDIJOF-FBSOJOH ৽ଔ࠾༻ ৽ଔΤϯδχΞืूϙδγϣϯ
.JDSPTFSWJDFT 1MBUGPSN *ΠϯλʔϯγοϓͷΈืू
64 .BDIJOF-FBSOJOH 4JUF3FMJBCJMJUZ .JDSPTFSWJDFT 1MBUGPSN 8FC1MBUGPSN 4JUF3FMJBCJMJUZ &.HS %JHJUBM.BSLFUJOH
4QFDJBMJTU த్࠾༻ த్࠾༻ืूϙδγϣϯ *5#13 .HS 4JUF3FMJBCJMJUZ &.HS