シャドーイング練習: System Design Mock Interview: Design WhatsApp - 動画で英語スピーキングを学ぶ

練習モードを選ぶ

各モードは全画面の練習ルームで開きます。

264 文

文が短すぎたり長すぎる場合は、Editをタップして調整してください。

can you tell me how you would design WhatsApp?

Hey everyone, welcome back to another Exponent System Design mock interview.

My name is Kevin Wei and on today's show we have Roshan.

Before we get started, Roshan, do you mind telling the audience just a little bit about yourself and then we'll get into the mock interview?

Of course, Kevin.

Thank you.

Thank you so much for inviting me to this interview.

My name is like roshan i work as a sd2 at amazon

and yeah i spend most of my time building highly scalable systems

and you know uh creating impact in the lives of billions of people excellent thank you roshan

so let's do this question today can you tell me how you would design whatsapp yeah of course

so maybe like usually whenever we start the system designs first we start with collecting all the requirements

so Kevin what I'm gonna do is first you know try to get all the requirements

that the WhatsApp application could have because today like you know everybody's aware of WhatsApp

and it has like a bunch of features

but in the interest of time maybe we will be developing only some you know essential features is that fine Kevin?

Yeah that sounds good maybe you can tell me what these essential features what the priorities might be sure

so yeah maybe i'll start writing it down

so first thing like maybe we can start with very basic

so today whatsapp is used uh you know to send

and receive messages uh you know um between um two peers

so yeah this is one of the most basic feature where

two person like one person one person can you know send

and receive message to another person

so this is like a you know the minimum feature

that any typical chat application should have the second uh one could be you know a group messaging where a member

would create a group they will add some x number of people in the group

and whenever the person sends the message all the other parties in the group will receive it

and everybody will know that who has sent that message

so that's the second feature that we may have

and we can also have like once the message has been

sent you know um the other person would be really interested

to know whether it has been delivered second one would be

like rather it should there should be three things three receipts basically the sent receipt the delivered receipt

and the read receipt so maybe in case

if we want we can add that as well sent delivered and read receipt

so yeah this could be another feature um so kevin anything uh

so far like on these three features do you want to add something yeah this makes sense sending

and receiving messages this seems like p0

group messaging is interesting and sending delivering read receipts just to let the user know

that the app is working and that everything that they're doing is working as planned

i think we can stick with this for now and then

if we want to add more features we can

if we have time okay so now maybe i'll just ask some more clarifying questions on this uh the first feature

so when we say messages

so are we talking only about the text messages are uh

you know the the person can send documents images videos um is that feature available in the first version

yeah great question let's assume that most of the messages are going to be text or images okay okay

okay got it so i think uh yeah so maybe i let me add some uh title in here

whatsapp system requirements okay

so now uh maybe i'll proceed with the you know the

design of like what what is going through my mind so um let me let me draw some icons in here

nice i love it some visuals for us yeah so let's say like this is the user a

and we also have another user in here B.

Okay.

So now we need some kind of a protocol where A can communicate to B

but in a more secured way so that we don't get, you know, like some third party should not be able to, you know, decode the message and,

you know, read it because that would be a pretty awful experience for the user B.

like to the to both the users for that matter.

So we need some kind of a protocol where we can you know establish

that communication and I think we have something called web sockets.

So web sockets was specially designed for peer-to-peer communication.

So maybe I can quickly write that down one second.

um web sockets however uh just using web sockets won't work uh how would like how would my system know

that you know this uh this user b is present in

some network like b should b would be connected to some network right

so we need some kind of a you know database where we know

to which box is a connected to and

which box is b connected to

so let me just change this a bit let me put this here

and let me remove this line over here and yeah

so let's say maybe i can here okay so now let me see

okay so we need api gateway

okay

so basically what would happen is this you know this phone

this this user a is connected to api gateway one

and there's another user b which is connected to ap api gateway two Now, maybe we can have a microservice.

Let me put that here.

And we can call it as a session service.

One second.

Session microservice.

Basically, the responsibility of session microservice is to connect with a DB.

And it will tell us which user is connected to which box or which API gateway.

So, maybe I think, yeah.

So say suppose when the message

when A wants to send a message to B it would first go to API Gateway

and then connect to this microservice

and this microservice will fetch you know the box number of the user B

and then the session microservice will connect back to API Gateway number 2

and from there it will go to API Gateway to be

and the protocol that we will be using here is web sockets maybe i can write that down real quick

so this is how i think i would be able to you know send

and receive messages and i think this solution is also pretty scalable how are we going to ensure

that is maybe the type of database that we will maintain here is maybe NoSQL database.

We can use that for scalability purposes as well.

Yeah, NoSQL like maybe DynamoDB.

So why would you use NoSQL here rather than a relational or SQL database?

As in when the number of users grows, we want the systems to scale, right?

And NoSQL are one of the most of the time why we use no sequels

100

because it helps us during scaling and

101

we don't have to stick to one schema the schema could be pretty dynamic

102

so yeah okay for scalability yeah it makes sense yeah

103

and maybe i think i can quickly write that format as well

104

so for now i'm just thinking of a key value pair so it could have something like um the key

105

and yeah so the key could be the you know um user

106

and the value could be the the box uh or the gateway

107

so by that by that way we'll be able to identify

108

which user is in which box so yeah so

109

that would only this this problem this particular design will solve the first requirement

110

but we still have not addressed the images thing

111

so usually for static content we have s3

112

that is the most preferred feature like you know most preferred option

113

so maybe i i think i would stick to that

114

so maybe i think when

115

when user wants to send a image right the only difference between text

116

and image here is um by the way we are not planning to store this uh text anyway right

117

because as soon as uh as soon as this um text

118

reaches the session microservice we are only going to get the user name

119

and their box number

120

and then send the text message as it is we don't

121

need to store this text anyway however in case of images i'm thinking would it make sense to store it um

122

we we can say that we don't want to store it just

123

because we don't store text

124

so we also don't want to store images for privacy yeah privacy purposes okay

125

so yeah in

126

that case then i think we are good with first requirement

127

maybe before we go to second requirement let's complete the third one So send, delivered and released.

128

Now, as soon as, you know, this user A has sent a message,

129

like as soon as a message is received by the session microservice, what we can also do is immediately send a response back,

130

like an acknowledgement to user a

131

that your message has been received by us this session microservice can do

132

that so maybe what we can also do is like let me put

133

that real quick okay and the text here would be a read um message um send acknowledgement

134

okay and maybe let me put some numbers as well

135

so basically the the number one here indicates uh the first part of it

136

because the message has been sent now once the session microservice has got the box number of the user b

137

and once it has sent that message to b

138

what we can do is it can send back an acknowledgement to session microservice

139

so maybe let me put it this way um api gateway

140

and from here to session microservice and let me add some numbers here

141

so this is more of a message received acknowledgement and

142

And what I can also do here is I can add another text in here.

143

It says message delivered acknowledgement.

144

Okay.

145

So just to summarize, user A has sent a message to APA Gateway.

146

APA Gateway would have communicated to session microservice.

147

And as soon as the session microservice has got a message from A, it will send a message sent acknowledgement.

148

That is point number one.

149

and session microservice would interact with DB.

150

It would get the box number and it knows exactly where user B is.

151

So it would go to API gateway 2 and API like you know this message will get redirected to B.

152

Now as soon as the message is you know received by B, B can again send a acknowledgement.

153

So that is point number 2 and as soon as session microservice gets that acknowledgement

154

it will send a message delivered acknowledgement to A.

155

So, you know, sent has been covered, delivered has been covered.

156

Now read receipt.

157

So the thing is, as soon as the user B opens the message, we can do exactly similar thing what we did for, you know, point number two.

158

So maybe I think I can draw another box in here.

159

So what happens here when user B reads the message?

160

yeah message read acknowledgement yeah

161

so now once the message has been read we need to you know somehow indicate

162

that to user a right

163

so what we can do is again we can use the

164

typical feature of what whatsapp is already averaging like maybe changing the color of the double ticks

165

so as soon as session microservice communicates that to a the the double ticks might be changed to double blue ticks.

166

So that would be point number four, I believe.

167

This is point number four.

168

Yeah, that makes sense.

169

Cool, yeah, that makes sense.

170

So this makes sense for peer-to-peer texts and images.

171

The texts and images are not stored.

172

They're just transmitted through the network.

173

We have sent, delivered, and read receipts.

174

Can we cover the group messages here?

175

So now let us talk about the group messaging.

176

So what essentially are we trying to achieve here is step number one, some user will create a group.

177

Step number two, the user will add some X number of people.

178

By the way, I forgot to ask this Kevin, how many, what is the maximum number of people who can be in the group?

179

yeah for simplicity let's say 10 people okay maybe I think I can write

180

that down real quick max 10 members yeah

181

and feel free to tell me

182

if in the future we want to change this to like 99

183

or 999 how that would change so maybe I think I can write

184

that down um should be scalable to more than 100 plus members okay got it

185

so maybe i'll think i'll start with 10 first

186

and then let's see how we can uh make it more scalable kevin so yeah

187

so maybe i'm just thinking of how i can tweak the existing design a bit to start supporting the group messaging.

188

So here if you see when a message was sent,

189

we had this key value pair with only one user and only one box.

190

So what if we have a list of a map?

191

So whenever a group will get created.

192

So what would happen is a list will get created with

193

them with a map having the key as the user and its associated box number.

194

So by that way, what would happen is whenever a message is sent,

195

the microservice or the messaging microservice will iterate over each of the elements in the list

196

or rather the the key value pair it will check the you know the the you know the user name

197

and its associated box number and it will exactly follow all the procedures

198

that we have created for one does that make sense kevin yep that makes sense How can we ensure scalability here?

199

Okay, the problem here is that when we are sending a message, we don't want to send the message in sequential order.

200

We can actually parallel, you know, parallelize things here.

201

We can ensure that, like, say suppose there are 100 members in the list, we can separate them in a batch of 10 or 20.

202

And we can have a main thread who will initiate a message, you know, who will initiate the group messaging, but the child thread will,

203

you know, equally distribute the number of items in the list and then send it to them parallel. By that way,

204

the other set of members in the list do not have to wait until the first set of members are done.

205

Got it.

206

Cool.

207

All right.

208

That takes our group messaging.

209

I have one more question before we wrap things up.

210

So you have user A here connected to API Gateway 1, you have user B connected to API Gateway 2.

211

If we have user C, where would user C be connected to?

212

Is it to their own API gateway or would they share a gateway?

213

No, they can share their gateway.

214

It's not mandatory that they need to have their own gateway can definitely share the gateways okay

215

and how would the how are the users assigned to the gateways is it uh randomly

216

or is there some structure there uh well for now i don't think we need any structure

217

but yeah they will be assigned randomly given okay got it

218

cool thanks roshan uh i think i understand where you're getting with this uh you somehow want to ensure

219

that not all the users end up in getting connected to one gateway right right

220

so i think yeah in order to solve

221

that problem what we can do is we can have a

222

limit on the number of users getting connected to a gateway

223

so by that as soon as that

224

that count reaches the max count we know

225

that okay the gateway is fully occupied now we have to move to the next gateway yeah

226

that makes sense um yeah thanks for clarifying

227

that point for me all right cool um thanks rishan for

228

your time i have a few pieces of pieces of feedback here

229

that i'd love to point out

230

and then i'd love it to turn it back to you

231

maybe you can give some advice to the audience as well yeah

232

so you did a really good job prioritizing the features in

233

the beginning figuring out what should be in the mvp you know whatsapp is a huge app

234

and you you were able to pick the ones that are most important to discuss for the MVP.

235

I think you also did a really good job of distinguishing some of the ambiguous parts of the question.

236

So, like sending and receiving messages, this could be anything from text to photos to videos to audio messages.

237

And you did a good job pointing out that ambiguity and helping to hone in the scope there around text and images.

238

And the last point I wanted to make is

239

that you did a really good job checking in with me and figuring out what questions I was trying to ask you, not only at the end there, but also in the beginning.

240

After you finished talking about sending and receiving messages, you didn't just move on to the next part of the question.

241

You paused and you asked me if I had any questions before moving on.

242

So I really appreciate your mock interview here.

243

And before we wrap things up, do you have any pieces of advice for the audience if they have their own system design interview?

244

yeah of course thank you kevin thank you for acknowledging this mock interview

245

and for people who are you know watching this video right

246

now um maybe i think i would just summarize of whatever

247

kevin has said just try to ask as many clarifying questions

248

that you have

249

because uh the only person to whom you can talk during the interview is your interviewer

250

so they are the ones who who will who are you who are here to help you out

251

so maybe you can ask as much clarifying questions

252

and whenever you are designing something so whatever is going through your mind

253

or whatever you are drawing just uh you know call it out loud

254

so that the interviewer knows what is going through your mind

255

and in case if you make any mistakes they will immediately stop you

256

or you know they might tell you to uh you know tweak your solution a bit

257

so yeah the key thing here is during the system design interview it's really really important

258

that you keep talking to your interviewer instead of you know just keeping it with yourself

259

so yeah makes sense thank you roshan thanks for the audience for watching

260

and good luck with your upcoming system design mock interview thanks

261

so much for watching don't forget to hit the like

262

and subscribe buttons below to let us know that this video is valuable for you.

263

And of course, check out hundreds more videos just like this at tryexponent.com.

264

Thanks for watching and good luck on your upcoming interview.

アプリをダウンロード

流暢に話すために必要なすべて

AI発音採点文ごとに採点

IPA練習すべての音をマスター

単語学習自分の単語帳を作る

単語ゲーム遊びながら覚える

QRをスキャン

Download on theApp Store Get it onGoogle Play

同じチャンネル

FOR YOU

このレッスンの語彙とスピーキングのポイント

このC1レベルのスピーキングレッスンは、動画「System Design Mock Interview: Design WhatsApp」を教材にしています。繰り返し出てくる語は次のとおりです：message, user, gateway, microservice, send。この動画には、シャドーイング用の文が264文、単語が3264語あります。音声の長さは22:22です。話す速さは1分あたり約146語で安定しており、シャドーイングしやすいペースです。単語の86%は英語の頻出3,000語に含まれます。残りは練習の前に確認しておきましょう。

この動画の重要語彙

動画の中で特に難しい単語15語を、発音と意味つきで紹介します。

単語	発音	意味
acknowledgment 名詞	/əkˈnɑ.lɪd͡ʒ.mənt/	承認, 承知
receipt 名詞	/ɹɪˈsiːt/	受け取り
socket 名詞	/ˈsɑkɪt/	コンセント
mock 動詞	/mɑk/	真似る
scalability 名詞	/ˌskeɪ.ləˈbɪl.ɪ.ti/	伸縮性, 拡張性
interviewer 名詞	/ˈɪntɚvjuɚ/	インタビュアー, 会見者
tick 名詞	/tɪk/	だに, ごさらぎ
schema 名詞	/ˈskiːmə/	スキーマ
summarize 動詞	/ˈsʌməˌɹaɪz/	要約する
assign 動詞	/əˈsaɪn/	割り当てる
initiate 動詞	/ɪˈnɪʃ.i.eɪt/	始める, 取り掛かる
decode 動詞	/dɪˈkəʊd/	解読する
prioritize 動詞	/pɹaɪˈɒɹ.ə.taɪz/	優先させる
ambiguous 形容詞	/æmˈbɪɡ.ju.əs/	曖昧な
fetch 動詞	/fɛt͡ʃ/	持ち帰る

動画に出てくる句動詞

単語	意味
point out 動詞	指す, 指し示す
check out 動詞	試験する, 検査する

この動画の文法

話し手がよく使っている文型を、動画の実際の表現とともに紹介します。

文型	動画での表現
条件文 if + 節, will/would + 動詞 — 条件とその結果	If we have user C, where would user · if you make any mistakes they will immediately
現在完了形 have/has + 過去分詞 — 過去の出来事が今も関係している	has sent · has been · has been delivered
受動態 be + 過去分詞 — 誰がするかより、何が起きるかに焦点を当てる	has been delivered · would be connected · is connected
関係詞節 who / which + 節 — 人や物について情報を加える	people who can · people who are · ones who who

注意したい発音

話し手はdon't, I'll, I'mなど、短縮形や弱形を28回使っています。聞こえたとおりの短い形で発音しましょう。

「sh」と「zh」の音: initiate /ɪˈnɪʃ.i.eɪt/, distinguish /dɪˈstɪŋ.ɡwɪʃ/, specially /ˈspɛʃəli/
長い単語（アクセントの位置に注意）: acknowledgment /əkˈnɑ.lɪd͡ʒ.mənt/, scalability /ˌskeɪ.ləˈbɪl.ɪ.ti/, initiate /ɪˈnɪʃ.i.eɪt/, ambiguity /æm.bɪˈɡjuː.ɪ.ti/, prioritize /pɹaɪˈɒɹ.ə.taɪz/

日本語話者が苦手な音：

/r/ と /l/ の区別 — /l/ は舌先を歯茎につけ、/r/ はどこにもつけない: clarify /ˈklæɹ.ɪ.faɪ/, randomly /ˈɹændm̩li/
/f/ — 「フ」ではなく、上の歯と下唇で出す: clarify /ˈklæɹ.ɪ.faɪ/, fetch /fɛt͡ʃ/
子音の連続 — 間に母音を入れない: acknowledgment /əkˈnɑ.lɪd͡ʒ.mənt/, scalable /ˈskeɪ.lə.bl̩/, clarify /ˈklæɹ.ɪ.faɪ/, scalability /ˌskeɪ.ləˈbɪl.ɪ.ti/, schema /ˈskiːmə/

この動画での練習方法

まず声を出さずに動画を最後まで聞き、知らない単語をメモします。
通常の速度で一文ずつシャドーイングし、話し手のリズムに合うまで繰り返します。
自分の声を録音して元の音声と比べます。acknowledgment, receipt, socketなどの単語に特に注意しましょう。

シャドーイングとは？英語上達に効果的な理由

シャドーイング（Shadowing）は、もともとプロの通訳者養成プログラムで開発された言語学習法で、多言語習得者として知られるDr. Alexander Arguelles によって広く普及されました。方法はシンプルですが非常に効果的：ネイティブスピーカーの英語を聞きながら、1〜2秒の遅延で声に出してすぐに繰り返す——まるで「影（shadow）」のように話者を追いかけます。文法ドリルや受動的なリスニングと異なり、シャドーイングは脳と口の筋肉が同時にリアルタイムで英語を処理・再現することを強制します。研究により、発音精度、抑揚、リズム、連音、リスニング力、そして会話の流暢さが大幅に向上することが確認されています。IELTSスピーキング対策や自然な英語コミュニケーションを目指す方に特におすすめです。

シャドーイングのやり方: ステップ別の完全ガイドを読む →

シャドーイング練習: System Design Mock Interview: Design WhatsApp - 動画で英語スピーキングを学ぶ

練習モードを選ぶ

アプリをダウンロード

同じチャンネル

Uber Data Engineering Mock Interview - Ride-Sharing Data Warehouse Schema

FULL Data Science Mock Interview | SQL, Statistics, Business

What is an API?

System Design Mock Interview: Design Facebook Messenger

How to Answer System Design Interview Questions (Complete Guide)

Senior Product Designer Mock Interview: App Critique of Yelp

Understanding TCP vs UDP

What REALLY Happens in Meta’s AI Product Sense Round

Anthropic Culture Fit | Behavioral Interview

おすすめ動画

Excellent 2 Years AWS DevOps Engineer Mock Interview with Questions and Answers | Cloud | Docker

Distributed Task Scheduler: System Design Interview (Stripe & Amazon Offers)

Mastering iOS Development: Mock Interview with Ramiro Neto

Small Talk Tips From a Matchmaker: Start Your Story in the MIDDLE!

Concurrency in Low-level Design Interviews w/ Staff Engineer

API Gateways in System Design Interviews w/ Ex-Meta Staff Engineer

.NET Mock Interview - Scenario Based - 7+ Years Experienced Candidate

[2026] TOEFL iBT Speaking Take an Interview | Step-by-Step Strategies #0008

How to Become a Great Software Developer — Best Advice from Top-Notch Engineers

01 - Course Introduction - Shopify App development with Remix