跟讀練習: I Coached an AWS Engineer Senior to Staff (Live) - 通過影片學習英語口語
選擇練習模式
每種模式都會開啟全屏練習室。
806 句
如果句子過短或過長,請點選 Edit 進行調整。
1
This is Ahmed.
2
He's an ex -AWS senior engineer with 12 years in the industry, and he's trying to break into staff.
3
Ahmed has the stories to get there.
4
A four -week migration with 20 partner teams, a disagreement with a principal engineer, two years leading a project end -to -end.
5
But, like a lot of smart engineers, he just doesn't know how to tell them in a way that sounds staff level.
6
I'm Steve Wynn.
7
I spent close to 20 years at Amazon, I mated to principal engineer, And as a bar raiser, I personally conducted nearly a thousand interviews.
8
The three things that I'll help Ahmed fix are making sure that he actually answers the questions being asked, telling stories that indicate that he's at the staff level instead of senior,
9
and communicating scope so his stories actually land with impact.
10
If you're in the Seattle area and want to be featured in an episode, a link to the application is in the description.
11
Let's get into it.
12
Ahmed, thanks for coming in today.
13
What's the thing that you're the most afraid about when it comes to interviewing?
14
Basically, when it comes to behavioral interviews, the most things that get me nervous
15
and worried about is how can I get the right answer for exactly what the interviewer is looking for.
16
So once the interviewer has a question in his mind, I have lots of situations that came to my mind from my past like 10 years.
17
But sometimes it feels like, does it really the things that the interviewer is looking for? or actually I should think about something else.
18
And sometimes I'm worried that I start to speak about the situation
19
and I have like my train of thought going very smoothly.
20
Then I figured out, oh, no, does that really answer your question or not?
21
So that's what's getting me nervous about the interview.
22
Okay, don't worry.
23
There's, you know, I can totally help with that.
24
Let's pretend it's a real interview.
25
And then we'll go through a bunch of questions.
26
And then after that, we'll debrief and then I can give you a plan for helping out with your problem.
27
Sound good?
28
Sounds good.
29
All right, let's do it.
30
So to start off, can you tell me about yourself, Ahmed?
31
Sure.
32
So my name is Ahmed.
33
I'm a senior software engineer.
34
I have been in the industry for almost like 12 years right now.
35
So I started my career back in Egypt.
36
I started like junior engineer, worked in a couple of local companies.
37
Then I joined AWS.
38
That was back in 2016 in South Africa.
39
I spent there about three years.
40
working at AWS and then moved here to Seattle just before COVID, like almost end of 2019.
41
And I was working at AWS Health Org.
42
So I got promoted there two times until I became like senior engineer.
43
And I just left Amazon a couple of months ago.
44
Okay.
45
How long were you a senior engineer for?
46
Almost three years.
47
And then how long have you been in Seattle for?
48
Sorry, was it 20 right before you said yeah it's almost like six years now okay all right
49
so let's just jump into the behavioral interview
50
so since you're interviewing for a staff position something like
51
that tell me about a time you had a conflict with a teammate
52
or a manager are you looking for the result
53
or the situation itself and how i dealt with it
54
so both okay true when you mentioned that i had like some stories
55
that came to my mind
56
and uh one of them i can pick is i have been working on a design
57
and i had a conflict with one of the engineers reviewing my design like a teammate basically
58
and we had like some discussion an argument
59
and um we went through this argument like by discussion with a senior engineer in the team
60
and this time i wasn't senior engineer i was like sd2 level engineer in amazon
61
and we worked through this discussion together we had some discussion with other senior engineer
62
and the team as a manager in the team until we resolved the conflict i can't go into some details
63
but just want to make sure that that and
64
that kind of situation that you're looking for how adult was
65
that what was the conflict the conflict basically i had
66
that i was working on a design and i came up with a solution
67
that can it's a little bit complicated but not not complicated
68
but require more works and building more components
69
because i saw some potential issues on the simple way
70
that everyone thought about and after research and experiment i figured out
71
that that will be a potential issue in the future
72
so i wanted to basically protect us from that
73
and came up with a little bit not complicated design
74
but more like building more components
75
and require more work to to get it i wrote the design document
76
and came up with evidence and some experiments from what I did and came up with different solutions.
77
One of them is a simple one, straightforward one, and the other one that required more work and reviewed with the team,
78
get some feedback, got basically some questions about, okay, how did you experiment that?
79
What was the issue that you expected from this solution?
80
Simple one, basically and
81
what your call basically because i came up with a different solution
82
and i have to come up with like one big one uh solution we discussed
83
that as a team i managed to convince most of the
84
team except one engineer who thought like okay no i don't think that's right
85
that was the senior engineer no he wasn't saying okay
86
so another mid -level engineer yeah at this time yeah
87
and uh like started to discuss that
88
and he went to do some homework basically to prove
89
that his point correct i want to do my homework to prove
90
that not to prove my
91
but more research okay maybe he's right maybe i was missing something will it be actually an issue
92
or i'm just over complicating the solution just to make my design more fancy
93
so i was convinced he was convinced we had an argument about
94
that so we decided like okay let's go to the senior engineer
95
and see what he thinks about
96
that we had a discussion with the senior engineer he was more aligned to my solution
97
because he was convinced that this actually, the simple one will be mainly like bite us in the future, basically, if we went with that.
98
Bite us in the future in terms of potential data loss, in terms of potential overhead and on -call and ops work.
99
So we had a discussion as a team with the senior engineer and the manager, and we decided like, okay, that's basically the route that we're going to take is the suggested solution.
100
So we went and implemented this way.
101
And during the release, one of the steps is basically to make a load test
102
and during the load test we figured out
103
that okay the other solution simple one would cause more problems
104
if we went with it because that the system would not behave in the way
105
that we want with this data log so as a result of
106
that we had like a real tangible uh example
107
that this solution the one that they came up
108
and convinced the team with is the one
109
that we should go with to the production we we did
110
that and we deployed and it was sustainable
111
and in terms of like after we released it was reliable and sustainable as we expected.
112
Okay.
113
Here's where almost everyone loses the interviewer.
114
The opening.
115
Most people start a story with a ton of background about the company, the team, and the technology involved.
116
It just takes way too long to get to what you did if you start this way.
117
You want to flip that around.
118
The first sentence out of your mouth should be the entire story compressed into one line if at all possible.
119
In my book, I call this the headline, and it's the most important sentence of your entire story.
120
Your goal is to name the problem you faced, the action you took, and why it mattered.
121
Something like, let me tell you about a time when I discovered our processing delays were caused by database locks, not the network issues everyone assumed,
122
and I fixed it by redesigning our update logic, which sped the system up by 16 times.
123
That one line confirms you understood the question and signpost exactly where you're taking them.
124
The trick is that you want to reverse engineer the interviewer's notes.
125
You hand them the sentence you want them writing down, then you pause and let them steer.
126
I break the full structure down in the book.
127
Back to it.
128
So for that example, how do you think you did?
129
Was it a good example?
130
I think, yeah.
131
I think it's answering.
132
Is it a staff level example?
133
It's staff level example in terms of, no, it's not.
134
No, I would say like some type of behavioral questions, I think maybe they are kind of the same for some levels.
135
In terms of like, that's why I ask the question is like, what actually you're looking for from the conflict level?
136
Like how big is the conflict should be?
137
Will it be like technical or will it be strategy conflict or will it be?
138
It can be any type of conflict.
139
Okay.
140
But I think for any question, your answer will tell you what level the other person is.
141
So if you told me like, hey, I don't know what level Ahmed is, but here's his answer to this question.
142
Tell me what level he is.
143
What level do you think that Ahmed would be?
144
I would say like senior.
145
Senior, even though you basically said, hey, the answer to this question is we went to my senior engineer and he agreed.
146
That's a good point because one of the things
147
that I think it makes sense to discuss is like how far I should go with my examples.
148
So immediately I...
149
Yeah, no, we can, we'll get there.
150
I think the big thing here is I think that's a really solid mid -level answer.
151
I actually think that it's not a good senior level answer and it's definitely not a staff level answer.
152
Is it because I mentioned that we went to discuss with a senior engineer?
153
That's one of the reasons as well.
154
Okay.
155
This episode is brought to you by Linear.
156
And they just published something that I think is worth talking about.
157
They opened with the new manifesto, issue tracking is dead.
158
And what that means is that the behaviors around it, like the manual logging of issues, the prioritization back and forth,
159
the grooming sessions, where half the room is just trying to figure out what's even in the backlog.
160
These things are a thing of the past with Linear.
161
I lived this nightmare at Amazon.
162
We had mechanisms for everything.
163
And don't get me wrong, I'm a big believer in mechanisms.
164
But more often than not, the process of tracking the work took as much energy as doing the work.
165
You'd come out of a sprint planning meeting and think, we just spent two hours so eight engineers could figure out what they already knew they needed to build.
166
What Linear has done is rebuilt the platform so that the stuff happens automatically.
167
Your code, your decisions, your customer feedback, all of that context all lives in one place.
168
And Linear actually works from that.
169
So when something comes in, the system already understands what it is and where it fits.
170
Nobody has to triage it manually.
171
If you're an engineering leader and you want your team building instead of administrating the work, check out Linear.
172
The link is in the description.
173
So I'll give you an example.
174
If I were to answer that same question, I have a story where I disagreed with the vice president.
175
So if I just, on the surface, if I told you about a time where I disagreed with the vice president, I didn't report into that person, but I disagreed with him,
176
and I had to interact with him and all of this other stuff, what level do you think I would be?
177
Principal level?
178
Yeah, right?
179
And so I do think that the level of the belligerents, if you will, the people that are involved in the story, it really matters.
180
If you disagreed with a principal engineer, that kind of makes me think like, oh, okay, well, this guy can actually push back against a principal engineer.
181
I think that's great.
182
Yeah.
183
Right?
184
If you're like, oh, there was a mid -level engineer that didn't like an aspect of my design, I think you're playing at mid -level engineer stakes just on the surface of it.
185
Right?
186
So that's one thing.
187
I think if I were you, I would think about like an example where there was a conflict
188
or a disagreement for as high of a level as you can think about for your experience.
189
Okay?
190
But let's just take this example and flesh it out.
191
So that's the first thing.
192
The second thing is, in this example, you turned out to be right.
193
I think what you want to do in a lot of
194
these situations is to talk about how the other person could have been right as well.
195
Stories where you're obviously right and they're obviously wrong, they don't tend to be good stories.
196
Because the reason people ask this question is they're going to go and hire a bunch of smart people.
197
And they're going to put all of these smart people in the room.
198
And then there are going to be, of course, there's going to be conflict there.
199
We used to have like a weekly principal and senior principal meeting in my organization at Amazon.
200
We didn't agree on anything, right?
201
That's just the natural state of things is there's going to be a lot of disagreement, right?
202
And so the question is like, can we make progress despite the fact that there are disagreements?
203
Or is like everything a hill that you're going to die on, right?
204
And it shouldn't be.
205
One thing about your story is I would have loved to understand the choices here.
206
I think what you said is like one side is like, hey, it's a slightly more complicated solution that's more involved and it'll take longer.
207
And the other person was pushing back and said, do we really need that?
208
Can't we just do the simple, straightforward way?
209
I would have loved a little bit more detail there.
210
And I might have asked it in a follow up, which is just like, what was the consideration?
211
What was the thing that they were really pushing back on?
212
Like complicated versus not complicated isn't the right framing.
213
You had said something around data loss.
214
Like maybe I would expand into that.
215
What's one more layer down without going into crazy detail?
216
What was the actual trade -off?
217
What was the basis for the disagreement?
218
Basically is one of the solutions that will require more storing the old data
219
and compare against them rather than just go on the fly
220
and rely on basically that downstream that we're using that will sustain
221
and load you wanted to build like a version store at least like keep the old values yeah and then
222
he's basically like no let's just update everything
223
when we need to update yeah just trust the downstreams
224
and i think they can scale
225
and was it like a control plane of some sort yeah okay
226
so the way
227
that i would say it is i think you were you
228
were trying to you wanted to say really really high level with the disagreement
229
because you didn't want to get into all of the details
230
but i think that's a false dichotomy you can just go one layer down in terms of the details
231
so that we can frame the disagreement a little bit more without going into,
232
hey, he wanted to use consistent writes and I wanted to do like eventual consistency or whatever.
233
Like we don't have to go at that layer, but we can just go one layer down, right?
234
So you could have said something like, hey, we were working on a control plane system.
235
Clients call in and set their values.
236
I wanted a more robust system where we kept like the
237
old values in case they wanted to roll back maybe a little bit more guardrails
238
so they couldn't shoot themselves in the foot.
239
And my colleague said like, why would we waste our time with that?
240
We need to trust our clients more and they know what those values are.
241
If we get pulled into every update, that's just going to overload our team.
242
Why can't we just push something out that's a little more simple and then we can add more, layer on more complexity later if it's needed.
243
I think that makes it sound like, oh, OK, well, there's two rational people that have a real choice in front of them.
244
Nobody is right.
245
In a sense, you're like your colleague was also right.
246
Yeah.
247
Right.
248
And so whenever we talk about conflict, we have to talk about we just had a difference of opinion.
249
It wasn't like I was obviously right and he was obviously wrong or vice versa.
250
You also talked about escalating it to like a senior engineer.
251
That's the thing that was a yellow flag.
252
It's just like, why does everything need escalation?
253
And so I think part of me would have been like, hey, we tried to work it out on our own.
254
We weren't making any progress.
255
And so I was like, hey, let's just, how about we just go to our senior engineer?
256
You know, we can both make our case and then they can choose which one is right.
257
Instead, the way that you positioned it was like, oh, we went to the senior engineer and obviously he was on my side because we were thinking about the future, right?
258
So that's kind of a little bit more like, yeah, I'm definitely right.
259
This guy's not right.
260
I think it makes sense.
261
To some extent, I was trying to avoid this route and saying, like, it was wrong.
262
Because one of the things that I mentioned is I went to do my homework to make sure, like, I'm not missing anything and he could have been right.
263
So that's what I was trying to do.
264
Maybe the words didn't serve me well here, but I was, my intention was trying to avoid this route because I just hate to say, like, I'm right and he's wrong.
265
But the thing is, even if you went to go do your homework, it also sounded like you went to go do your homework so that you could argue that your side was right.
266
Like maybe you could have done some homework
267
and then shared the result with the person that disagreed with you instead of just being like, hey, obviously this is correct.
268
And then I escalated to my senior engineer and he thought I was correct.
269
And then I did the homework and then, you know, I had results and data, right?
270
You could have just been like, hey, I tried to make my case.
271
I went and got some data.
272
I presented that data to him.
273
He still wasn't budging.
274
And then only as a last result did we escalate.
275
I think it's a good example for an SDE2.
276
I don't think it's a good senior level example, and I don't think it's a good staff level example.
277
But I think if you were going to keep this story, that's how I would position it.
278
Do you have an idea for an example that would be senior or staff?
279
Yeah, I think I have much better examples.
280
If you have an example where you push back, you have a disagreement with a principal engineer at Amazon,
281
I would so much more prefer that than one where you were wrestling with another mid -level engineer.
282
I think that's the first one that came to my mind.
283
Instead of you delivering that, let's workshop this story so we can start off on a good foot, right?
284
So can you give me the headline for this story with the principal engineer?
285
So basically the summary for it is was I was working on a big project.
286
I had an idea how we can came up with a solution
287
that we can protect basically our downstreams from taking more ownership on how they manage the data on their site.
288
So we can leverage our data that we consume from like our upstreams
289
so we can protect downtown streams
290
because we already know how they are suffering kind of from service like it was a legacy service
291
and they were basically suffering from some data load on their side and how they keep it consistent.
292
And this was like, we were in the same org.
293
So it was like a strategy level thing on the org
294
to how we can protect this service from being overloaded with data.
295
Yeah, so for this story, like let's start with the conflict.
296
Yeah, sure, yeah.
297
Again, we want to do just in time context.
298
Yeah.
299
I think I understand your side.
300
Let me tell you about a time where I thought
301
that my team was in a spot where we could do more for our downstream clients
302
so that they didn't have to worry about all the overhead of maintaining this data or whatever it was.
303
But my principal engineer thought...
304
Let them do it, basically.
305
Yes, because...
306
Because that's going to make our service a little bit complicated
307
that we're going to deal with some situations in the future
308
in terms of how data is stored in our site and processed it, that they are basically built for that their service built for
309
that they have the capability to do it like although they are maybe suffering from it a little bit
310
but if there let's fix it there instead of those on our side yeah
311
so i would say something like you know let me tell you about a time
312
when i disagreed with a principal engineer i thought
313
that my team could do more for downstream clients and
314
that we could manage a lot of the data
315
and the workloads for them and that would make them sort of happier with us.
316
They would have to do less.
317
But my principal engineer thought like, let them do that work.
318
That's part of their responsibilities.
319
You shouldn't actually be doing all of this stuff.
320
I don't even need to know the underlying like data architecture.
321
I don't need to understand what the processing is.
322
The interesting part is this disagreement that you had to work out with the principal engineer.
323
And you are a senior engineer at this time.
324
If that's the setup, if you could tell the rest of the story, I'd be like, yeah, this guy's a staff level engineer, right depending on on what you did yeah okay
325
so let's continue with the story so yeah we we discussed about
326
that and we came up with a strategy like okay let's let's not do it right now
327
so i was convinced from the principal engineer
328
that okay maybe that's not the right thing to focus on
329
right now although like i've been think like it basically for me it was a disagree
330
and comment and commit right so you i disagree disagree i disagreed with the sentiment
331
that okay let's let them do it let them fix it there
332
if they cannot do it they have to handle it it's not our responsibility basically to do it
333
and um i disagreed and commented that
334
so i adjusted my design to basically have the idea there
335
but it's not something
336
that we're gonna implement right now we moved with our design we implemented we deployed in production
337
and maybe a couple of months later we figured out
338
that they are suffering from okay and what was better why
339
because they didn't have the capability to handle the load
340
that we put on them okay
341
so part of this discussion is you were intercepting traffic from these downstreams yeah and
342
so you were like we're in a better spot we can
343
aggregate all of the load exactly we can handle it
344
and your principal engineer is like no just give them the
345
traffic they should be set up for it you said your thing
346
and then and then you disagreed and you committed And you're like, fine, we won't do it.
347
And then it turned out that you were correct.
348
Yeah, it turned out that it was much easier for us to do it, less expensive, and we can handle it much better because of our situation,
349
not just because of how, because basically the other service was legacy and they have to deal with other data storage issues.
350
On our side, it was much easier to do it.
351
And yeah, Calder Funds later ended up that we are actually implementing this solution.
352
Okay, so then you went back and you did it the right way.
353
Yeah.
354
Okay.
355
How satisfying of a story do you think that is, the one that you just told?
356
In terms of what?
357
It's fine in terms of...
358
The interviewer.
359
Do you think that he would like that story or not?
360
If we break it down, yeah, I think it was argument that I stood up to my idea
361
and I tried to showcase from what I think is the right thing to do.
362
After, like, discussion with the principal engineer, I had to disagree and commit, so So I'm not really pushing to my,
363
so I'm not creating like bad environment and try to push from.
364
So basically having, if we think about that from Amazon principles, that's have a backbone and disagree and commit.
365
Yes.
366
And at the end, be right a lot.
367
Well, it's more of a be right a lot for the principal engineer, not for you, right?
368
Let's think about it this way.
369
When you disagreed and committed and then did it the way that the principal engineer did,
370
What would, I think, make this story really shine is that you turned it into a reversible decision.
371
When you guys figured out that you needed to actually take the load
372
and so you re -architected things to take the load from your downstreams, was it relatively easy to do or did it take, like, was that a really big effort?
373
It was another project that me, myself, like, they used my design, but me, myself, didn't implement it.
374
So I wouldn't say, like, relatively easy, but compared to the other ever yeah compared to the other effort
375
that had to be done on the downstream yeah so
376
because it was easier what i would say is at the
377
end of the day the principal engineer was the one
378
that was going to make the call and
379
so you know i had to disagree and commit after i said my piece
380
but what i did was that
381
when we were developing the solution we made it a little bit easier to reverse course
382
if afterwards the traffic had to come to us you probably did some things
383
that you were just kind of like anticipating like, hey, maybe if we need to reverse this decision, it'll make it slightly easier.
384
So when we implemented it, we made it slightly easier for it to come back to us.
385
And so when it turned out that I was ultimately right, two things here.
386
One is that you should make sure to not rub it in the principal engineer's face.
387
I think it's a good idea to take it back to the principal engineer.
388
So maybe in the future, if you have something similar, just be like, hey, turns out when we made that call, it wasn't the right one.
389
Then you can just be like, did we make the right choice or not?
390
I think it's a really senior level thing to do to understand
391
that we are all operating in this environment where there's some risks associated with our choices.
392
We don't, it's not a game of perfect information.
393
We're not playing chess where everything is out on the board.
394
It's more like we're playing poker and there's some things that we see and some things that we don't see.
395
So we shouldn't fall into this trap of looking only at the outcome.
396
But I do think that closing the loop with the principal engineer and saying something like, hey, there are some things that we saw and some things that we didn't see.
397
Here's how it worked out.
398
Did we make the right call knowing what we knew back then, right?
399
And are there any lessons for us to take it forward?
400
I think if you were able to do that in your story, that really makes the story pop.
401
Does that make sense?
402
Was there a moment where you went back to the principal engineer and you were like, what the heck?
403
No, actually it was the call from the principal engineer that he came back to me and said like, okay, I think we should have done it this way.
404
Great.
405
Okay, so that was left out of your story, right?
406
Yeah, maybe I didn't mention that.
407
Yeah, so again, I think the way that you position that is really important, right?
408
Just to really punch up your story, we're going to say like, hey, during our implementation, we made it a little bit easier to reverse course,
409
but I still committed and we did it the way that the principal engineer wanted.
410
And then you can be like, hey, turns out I was right.
411
It's OK.
412
I didn't rub it in his face.
413
He came back to us.
414
He came back to me.
415
Yeah.
416
And then we talked.
417
What was that?
418
What was that conversation like?
419
No, it was like smooth conversation.
420
I had a very good relation with the principal engineer.
421
So it was very smooth.
422
It was like, okay, it turned out that we want to have this component on our side, so we will do it.
423
Yeah, and because you had done a little bit of extra work during the implementation to make it reversible, it was totally fine.
424
If we can kind of like put those pieces together, I think those would be, that's a senior to staff, staff level example.
425
You feel better about the answer to that one?
426
Yeah.
427
Okay, good.
428
Ahmed just told a much stronger conflict story.
429
And the reason it's stronger is that he disagreed with someone above his level.
430
But he left out the single best detail in the whole thing.
431
After the disagreement, the principal came back to him and said, Ahmed, you were right, and that they should have gone with his approach.
432
That's the best detail, and Ahmed almost completely dropped it.
433
Here's why that's gold.
434
There's a huge difference between you saying you were right
435
and the other person coming back and telling you that you were right.
436
If you say it, you kind of sound arrogant.
437
and the interviewer starts wondering whether you're the difficult one to work with.
438
When the principal admits it on their own, it's exactly the kind of signal that levels you up, because you showed the conviction to push back on someone above you, but still commit to the decision.
439
The mistake Ahmed made is one I see constantly.
440
People think that the disagreement itself is a story, so they stop at, we ended up going with this plan, or we agreed to disagree.
441
But the outcome is the part the interviewer writes down.
442
If your judgment got validated, especially from above, you have to put that in the story.
443
Back to it.
444
So tell me about a time you delivered under really intense time constraints.
445
Yeah.
446
Remind me was one of the biggest releases that I have done.
447
So we had a big project to deliver.
448
And basically it was that the complexity of the delivery wasn't just
449
that we are releasing in production it was the complexity
450
that we're replacing one of the components that exists already with another component
451
so like not components basically like a big multiple components
452
so it was like kind of migration left and shift kind of things
453
and the complexity came with how the ecosystem will behave once we have the migration
454
so you have lots of up streams and down streams
455
and to align all the timelines across these multiple teams was a challenging part
456
and it was tight on time
457
that we had lots of things to finish on our side
458
and I was the lead for the project
459
and it was actually tight to make sure that all the upstreams
460
and downstreams are actually going to deliver on the time because
461
if one of them slept the whole project slept okay
462
so I had to work with the team to basically focus
463
on focus on the what's left to do to get our piece delivered
464
and work with other teams to see on their side what's left on their side to deliver on their time
465
so we can make the end -to -end story successful
466
so this time we were working on the normal scrum process in my team
467
so I had the decision with my manager to switch
468
that to like let's move it to a campaign board let's move it let's move
469
So let's list all the things
470
that we have on our site to get it done
471
and work with the team to make sure like who is responsible for that component so we can move fast.
472
So spoke with one engineer.
473
That's your area.
474
You are special in this one.
475
Let's get it done.
476
Another engineer.
477
Let's get it done.
478
And my piece.
479
Let's get it done.
480
At the same time, I worked with TPM
481
so we can align across all other teams that are consuming our APIs
482
so they can deliver their piece in time and making sure that how we can facilitate you, how basically we can solve things for you,
483
maybe make the onboarding much easier to our APIs.
484
Also work with other team that was like kind of downstream as well to our data, our service to make sure that they are right on time.
485
So what can we do on our site?
486
Let's help you with this implementation.
487
We are going to make sure that we provide you like a client much easier to onboard to.
488
And we're going to explain everything to you if we want to, like, if you want to, some of us working with you to deliver, we can offer that as well.
489
I also worked with, like, the stakeholders during, like, they had a backbash to go through all the things
490
that left on our site to work to them, like, okay, what are the, move more, like, strategic way is how we can make sure that what we deliver, what we're working on is B0 for you.
491
We prioritized everything with them and the product manager
492
and our SDM to get what's really important for them to get delivered first.
493
And if something is not really important that they thought it's B0, it's non -negotiable, we worked with them.
494
We had a couple of features that we worked with them to say like, okay, that's not really a B0.
495
We can't go without it.
496
Let's move fast for now.
497
So we had like lots of conversation with them to make sure
498
that we deliver something valuable
499
and at the same time we're not just I'm not saying wasting
500
but we're not working on the most important things right now
501
so I had to manage all these parties together to make sure
502
that we deliver on time and the timeline was very tight
503
because it has been moved multiple times because of us sometimes
504
because of other teams sometimes but it was like a set date
505
that we had to meet I think the most important things
506
that I have been working on is two main things the one I mentioned
507
that we're working on the right priorities
508
and the other one is how to make sure that
509
that each team is like team members are more focused on
510
one thing not just working across the board with different things
511
so helping like one engineer working for example just for one example working on operational excellence
512
so they are not just focusing on multiple things at the same time just focus hyper focus on one thing
513
and get it done so that's how I managed the deadline having managed all the team members
514
and all the assets that we have to meet the deadline.
515
And we ended up to actually have like, like meet the deadline and we had like a whole room thing to release and it was successful at the end.
516
We managed to release in the time, the first main component, and then we broke it down the other component to deliver that wasn't that tight deadline.
517
Okay.
518
How do you think you did with that story?
519
From your feedback before, I believe, I strongly believe that this is like a staff level because it's working across multiple teams,
520
multiple personas, and making sure that everything is aligned together to get the things delivered.
521
I try my best to make sure that I don't go into details and summarize what actually the actions have been.
522
I have talked to each stream, work stream, to get it done.
523
So yeah, I think it's...
524
Okay, so I would agree with you if the numbers make sense and the scope makes sense.
525
I think you might've gone a little too high level with the story.
526
And so whenever you are asked like about delivery and tight deadlines, you want to think about it like a movie.
527
Like how impossible of a situation are you in right now?
528
If you're able to get out of that situation, you'll be like, dang, that guy's pretty good.
529
So one thing is, I think you want to give a sense of how crazy the timeline is, how aggressive is the deadline, right?
530
So when we joined the movie, Ahmed had how many weeks or months to deliver?
531
It was a long -running project, but I would say like from four to six weeks.
532
Okay, so you're like, let's just say it's four weeks.
533
So Ahmed was put in the situation where he had to deliver something in four weeks.
534
Can you point out like the incongruity or the asymmetry between four weeks and like what you had to deliver?
535
Why was it difficult?
536
Yeah, okay.
537
I thought I mentioned that in my story, but it was difficult basically because we had lots of work streams that we need to do.
538
Yeah, so lots of work streams can be like two or one or like 50, right?
539
Yeah.
540
So I think one thing that's difficult for us is like it's difficult to communicate how big something is.
541
We start to run out of like words.
542
We can be like big, like really big, like gigantic, enormous.
543
Those are, people could say that forever.
544
So what I think would be helpful is to give like a relative size comparator.
545
And so is there something where the number is just like kind of obscene for a month?
546
Maybe you can say something like, hey, we had to deliver something in four weeks
547
that we delivered something similar to that last year and it took us four months, right?
548
Then I'll be like, dang, like you're being asked to move like 400 % faster the
549
number of work streams is difficult you could it's like story points they don't mean anything i can tell you
550
that something is like seven story points and you're like i don't know
551
that depends on the team yeah right and
552
so i was like is there a way for you to communicate the size
553
and scope of this project yeah i think that's a tricky part is how can you communicate the size
554
and of the project without going into
555
so what i would say is something like um the number of partners yeah
556
so number of upstreams and downstream so like the number of dependencies that are there we had to work with 20 teams.
557
I'm just picking a number.
558
No, actually it was close to too many teams.
559
Okay, good.
560
Yeah, so if it's like anything where you have to get 20 teams to do something for you
561
and they're in the dependency chain, a month is not enough time.
562
Like a month is 30 -something days.
563
Are you going to really like work with that many dependencies in 30 days?
564
Like that's pretty nuts, right?
565
So like let's add that critical bit of detail.
566
Like how many teams were there?
567
We'll use 20.
568
So yeah, and then I think how many P0s were there that you had to push back on?
569
P0s, you mean like they set a P0 and we had to push back to get P1 or P2s?
570
Well, no. So there's the initial list of P0s and P1s.
571
Yeah.
572
And then there's the actual list of P0s and P1s that you delivered with.
573
And usually those aren't the same thing.
574
Yeah. So like.
575
Yeah, we had to push like the numbers.
576
We had to push about three or four, but they are main features.
577
So they would take us at least four weeks to implement.
578
Okay.
579
There are three or four big.
580
And how many did you launch with?
581
A lot like 20 plus okay you had to cut some pretty significant scope the thing is is how
582
because as you mentioned like i can see we had to push one feature
583
but this feature is big
584
that will take us at least a month to deliver can you give like what was
585
that feature we had to implement kind of the data dedubing on one of our components dedubing
586
and undedubing so that will will include like lots of data migration
587
and data merging and so some of the b0s
588
that we had to push wasn't that big
589
but we have been focusing on the things
590
that we actually no we cannot do it in four weeks
591
so that's the way
592
that i've been saying like we're trying to thought about the strategic ways what is the actual b0s
593
that is meaningful for them it's we cannot that's non -negotiable we cannot release without them
594
or what are the features that actually they thought that it's b0
595
but after some thoughts from us
596
and some discussion with them is no it's not actually b0
597
or it's not actually b0 on the way
598
that has been written as a b0
599
so yeah of course we can just uh compromise on some of the features there
600
so that's the strategic way that we
601
so what i would say is like just going back to
602
setting the stakes yeah like why is this a big project right so i would say something like Like,
603
hey, let me tell you about this time when we had to deliver on something in four weeks.
604
That's really ambitious because there were 20 teams that were involved upstream and downstream.
605
And so something similar that we had built last year took us double the time or three times the amount of time.
606
And so we were being asked to do it in four weeks.
607
And that timeline was not negotiable.
608
They weren't going to move the timeline back.
609
On top of that, the list of P0s, right, the must -haves, was just way longer than it should have been.
610
They wanted everything.
611
So we were in the spot where there was a high amount of complexity due to the number of partners
612
that we had.
613
The timeline was not movable, and they wanted a whole bunch of stuff.
614
And we had to like sit there and educate them and really push back on what was there.
615
If that was true, then I think that, yeah, that's like the senior engineer staff level example, right?
616
But we have to anchor on relative measures of size.
617
Yeah.
618
So if you had said something like that, it would be like, okay, interesting.
619
If he is, if he found a way to like, to navigate that, that seems to me like a staff level engineer.
620
Right.
621
And so I think that was maybe missing.
622
So just like a little bit of just enough more context to understand like what the stakes were.
623
Okay.
624
Then I would say something like, um, I think you did a good job of like coming in and talking about what you did.
625
So, hey, we were doing like this big heavyweight scrum process where we like groomed the backlog
626
and we had like a retrospective and then we had like our backlog meetings.
627
But because the timeline was crazy, I came in and it was like martial law.
628
Nope, we're going to Kanban.
629
Kanban is basically one big priority queue we always just take from the priority queue.
630
Another thing that we did was like, we didn't have time for randomization.
631
So I said, Steve, you are in charge of this part.
632
Ahmed, you're doing this.
633
I'm owning this part.
634
If any of these things are blocked or whatever, immediately escalate to me.
635
So I like reduce randomization for our team.
636
Then I went back to our stakeholders and basically said, hey, this cannot be the list of P zeros and must -haves.
637
And so what I think is for like senior level people, you really just have to do like basic project management stuff.
638
You either move the date, well, the date's not movable.
639
You either add resources.
640
Typically when you add resources, there's some sort of like on -ramp period.
641
Well, if you can't add resources in time and you can't move the date, the only thing is to cut scope, right?
642
So I went back to them and I was basically like, hey, this date is at risk.
643
Is this stuff really, really important?
644
Do we actually need to de -dupe or is
645
that a nice to have and we can fix it later after this date?
646
And so I was able to convince the stakeholders that like three or four or five of these P zeros, we could triage and we can move on to the, we can do a fast follow.
647
I think that would be a pretty good example.
648
I think one last thing is what was the project?
649
If you could have just at a high level said like, Hey, we were trying to do X.
650
Yeah.
651
Like that can maybe communicate a little bit more of the complexity.
652
What was this project about?
653
It was basically a migration for an old system that has been running around for lots of years.
654
So we had to migrate it to a new system, but the new systems that we built, basically we like, we built it from ground up. So yeah.
655
But what did it do?
656
It's a ecosystem for managing the large scale events.
657
That was the project.
658
Okay.
659
Why was the deadline there?
660
The deadline, because the system has been suffering, the old system has been suffering a lot, and that was like a deadline that we came up,
661
that leadership, so it was coming from the leadership that you have to cut it down at this date.
662
So it wasn't negotiable as well because it has been pushed multiple times from other teams,
663
one of the town's teams that has been deleting us because of multiple changes in the strategy on this side.
664
So what I would say is we were building a new system
665
that handled large -scale event incident reporting leadership was not going to move the date anymore
666
because we had kicked the can down the road for a couple of years now
667
and they said no no it has to be done in four weeks
668
and no we're not going to extend the deadline
669
and we were in a weird spot
670
because we weren't evolving a system we had to lift
671
and shift and build in like a net new system
672
that had to work from day one
673
and we had a crazy deadline of four weeks
674
that wasn't going to get pushed and so like giving enough of
675
that context there it'd be like okay cool like this is a real deadline yeah like
676
if you didn't hit the deadline it would have looked really bad for you and your team, right?
677
So that also increases the stakes.
678
Makes sense.
679
Cool.
680
And then when it comes to delivery stuff, the results, we want to do a little bit more than just be like, we were able to deliver on time, right?
681
So if you can put in like, hey, we were able to deliver on time, it'd be good if you were like, hey, we were able to deliver even before the deadline, but also something like, you know,
682
there weren't any incidences or we were able to actually do the fast follows as a fast follow.
683
maybe you can add some measure of like client
684
or partner level satisfaction hey the 20 teams involved like they
685
gave us a pat on the back they were like i can't believe
686
that you finished this in time that sort of thing i think
687
that makes it work really well as well yeah cool makes sense all right
688
so you feeling better about that story yeah what's your best story
689
that you have i would say this one oh this is your crowning achievement yeah i would say
690
that because i had lots of details
691
and it has been like a long -running project for almost two years okay
692
so here's a trick it was a three -year project two years okay
693
so this like last little bit is part of like a
694
like a two -year epic like a two -year arc yeah
695
so the trick here is like we want to take like the biggest crowning achievements
696
and we want to turn that into multiple stories
697
if we can yeah that's what i'm trying to do usually
698
that's why you i think with more questions coming you will see
699
that i have been taking some examples from this project
700
so that's why i have been focusing so i had lots of deadlines
701
that was really pushing and it was i had to do lots of things to to manage
702
but this one the thing i like about this project is i had to play multiple roles
703
so i had to play senior engineers as a team take lead pushing people
704
and try to solve any complexity there
705
and the same time working with as a product manager
706
but working closely with a product manager that we have to advise them
707
and see what's the things
708
that we can solve better also working with a technical product manager to align with other teams
709
because i have lots of dependencies.
710
So that's the thing I like about this project is I had to play many hats.
711
Going back to the story that you just told, one last thing that would make this story pop a little bit is
712
if there was one last obstacle or challenge that put the project at risk.
713
Hey, I went in and we took Scrum away.
714
We're just doing Kanban.
715
I assigned responsibilities to every single person so they weren't getting randomized.
716
I went back to all of our stakeholders
717
and just really poked at some of the P0s and the P1s
718
and all of that stuff and everything looked like it was on track.
719
But then insert some sort of like challenge or obstacle at the end.
720
There was a really big problem when, and then what you can do is you can just take, you can take a problem that was already existing and you can basically be like,
721
hey, there was an assumption that didn't hold.
722
Yeah.
723
Or anything that's sort of like last minute to put the thing at risk, this is a good time to put it in there.
724
You don't want to fold that into, was there something about it that put the project at risk?
725
Yeah, actually there was.
726
And there was like one of the things that big features that we thought it's not required.
727
And it was actually required.
728
So we had multiple discussions with the stakeholders.
729
What was it?
730
Okay.
731
It was actually the one that we pushed, the P0 one, the deduping one.
732
So this one wasn't required.
733
And to come up with a better solution.
734
I had to work with a product manager to convince them
735
that it's not actually required and we're going to build something else for you
736
that we thought we can do it in this.
737
It was like maybe less than four weeks at this time.
738
So I came up with a solution to solve this issue in a more like, it's not the hacky way, but it's more like a manageable way.
739
So we can do it.
740
And I had to push with other engineers in the team.
741
We pushed hard to get it done before the deadline.
742
It was a couple of days before the deadline.
743
So we managed to do that.
744
We managed to come up with the idea, implement it, test it, and yeah, move it to production.
745
So yeah, so the way that I would say it is when we got, we finally got to the point where I was pushing back on some of the requirements,
746
some of these P zeros, and we were able to eliminate like three or four of them.
747
But then the PM came back and was basically like, no, for this deduplication issue, we can't push this.
748
This has to come back in.
749
And so I tried to argue with him, but then I realized, no, they're right.
750
If we don't dedupe, a lot of this effort is going to go to waste.
751
And so then you can tell a story about how you were able to pull that in, even though initially you had to cut scope, but then somehow the scope increased.
752
And so that makes the stakes a little bit higher as well.
753
And that makes it much more realistic.
754
Like, at the end of the day, like, all of this stuff is happening within a month.
755
And so, like, you're trying to push it out and then it comes back in.
756
Like, life's a lot messier than these stories that we tell, right?
757
So I think we can, like, introduce one wrinkle there.
758
Yeah, I think the tricky part about this kind of situation, there are lots of details there.
759
And it's kind of hard to pick exactly what's actually meaningful for the interviewer to listen.
760
Yeah, you have to, like, simplify.
761
So what it comes down to is like, I think for you, you summarize too high of a level.
762
Yeah.
763
And then you're like, okay, let's go into the details.
764
And then you go all the way down to like the nitty gritty details.
765
And we just have to go intermediate.
766
Yeah.
767
Right.
768
So every senior engineer, staff level engineer, every dev manager, they have all been in the situation where they're haggling over requirements.
769
Yeah.
770
Every single one of them.
771
You don't have to explain like the product manager pushed back and he actually said it's a P0.
772
You don't have to go into the details on that one.
773
Everybody that's a software developer, like in their soul, like they have been in that situation before.
774
Right.
775
So you can assume that they've been in that situation.
776
I think the situation that you're trying to surface is basically you can just say like, hey, there's some deduplication and undeduplication that needs to occur.
777
I realized that when I was pushing back on it and then we cut it out, I realized that we needed to pull it back in.
778
And it sucked because we were already on this time crunch.
779
And so I had to, and you can tell a mini story here, I had to like, luckily I didn't have to design it on the fly.
780
I already kind of knew what we had to do.
781
But even on this design, we had to cut scope on the design
782
and then just get it down to its bare essence so that we could get it across the finish line.
783
So then I think putting all of those pieces together, I think that's a pretty decent delivery story.
784
We have time for one more.
785
So maybe we can go this way.
786
What's a question that you are the most scared that they would ask that you don't have a story for?
787
Ahmed just told me the question he's the most afraid of, and that's where the best coaching in this whole session happens.
788
To see how I take him all the way through it, the full conversation is in the YouTube members cut or on my sub stack.
789
And that's a good place to step back and look at what happened across the whole interview.
790
Here's what stood out to me about Ahmed.
791
He was never short on material.
792
The experience was all there.
793
What was missing was knowing which stories to tell and how to tell them.
794
And that's the part almost nobody gets taught.
795
So here's the system that I'd leave you with.
796
It's the spine of the first half of my book, and it comes down to three pillars.
797
Pillar one, build the story so it's easy to follow.
798
Lead with the headline that gives the whole thing in about 15 seconds, then keep your core to about two minutes.
799
Pillar two, make sure the story is strong enough for the level that you're targeting, because a mid -level story will not get you a senior offer, no matter how well you tell it.
800
And pillar three, match the story to what the company actually values
801
because the same story can crush at a startup but fall flat at a big enterprise.
802
Get those three right and you'll stop hoping the interviewer connects the dots.
803
You hand them the dots already connected.
804
The full framework is in my book.
805
And Ahmed, thank you for being willing to do this on camera.
806
That took real guts and you leveled up right in front of everyone watching.
同頻道
FOR YOU
推薦影片
本課的詞彙與口語要點
這節口語課以影片“I Coached an AWS Engineer Senior to Staff”為素材。 這段影片共有 806 個句子、9927 個單詞可供跟讀。 講話部分時長為 49:37。 說話人語速較快,每分鐘約 200 個詞,會出現較多連讀和弱讀。 88% 的單詞屬於英語最常用的 3,000 詞,詞彙容易理解。
影片中的重點詞彙
影片中 15 個值得學習的單詞,附發音和釋義:
| 單詞 | 發音 | 釋義 |
|---|---|---|
| disagree 動詞 | /dɪsəˈɡɹi/ | 不同意, 反對 /反對 |
| deadline 名詞 | /ˈdɛdˌlaɪn/ | 截止日期, 限期 |
| component 名詞 | /kəmˈpoʊ.nənt/ | 元件, 部件 |
| interviewer 名詞 | /ˈɪntɚvjuɚ/ | 訪問者 /訪問者 |
| convince 動詞 | /kənˈvɪns/ | 說服 /說服 |
| scope 名詞 | /skəʊp/ | 範圍 /範圍 |
| implement 動詞 | /ˈɪmpləmənt/ | 實施 /實施, 執行 /執行 |
| complicated 形容詞 | /ˈkɑm.plɪˌkeɪ.tɪd/ | 複雜 /複雜 |
| homework 名詞 | /ˈhoʊmˌwɝk/ | 作業 /作業, 家庭作業 /家庭作業 |
| complexity 名詞 | /kəmˈplɛk.sɪ.ti/ | 複雜性 /複雜性 |
| timeline 名詞 | /ˈtaɪmˌlaɪn/ | 年表, 時間軸 /時間軸 |
| layer 名詞 | /ˈleɪ̯ɚ/ | 層 /層 |
| stake 名詞 | /steɪk/ | 木樁 /木樁 |
| migration 名詞 | /maɪˈɡɹeɪʃ(ə)n/ | 徙動 /徙動, 遷移 /遷移 |
| align 動詞 | /əˈlaɪn/ | 對齊 /對齊 |
影片中出現的短語動詞
| 單詞 | 釋義 |
|---|---|
| come up with 動詞 | 想出, 提出 |
| figure out 動詞 | 弄清楚 |
| turn out 動詞 | 原來 /原來 |
| come back 動詞 | 回來 /回來, 返回 |
| come up 動詞 | 上來 /上來 |
| end up 動詞 | 到達 |
| go through 動詞 | 通過 /通過 |
| check out 動詞 | 退房 |
值得反覆跟讀的句子
影片中簡短而完整的句子,可以直接用在日常對話裡:
- There's, you know, I can totally help with that.
- How long were you a senior engineer for?
- Can't we just do the simple, straightforward way?
- You feel better about the answer to that one?
- How do you think you did with that story?
影片中的語法
說話人最常用的結構,並附上影片中的原話:
| 結構 | 影片中的用法 |
|---|---|
| 現在完成進行時 have/has been + -ing — 以前開始、現在仍在繼續的動作 | have been working · have been focusing · i've been saying |
| 條件句 if + 從句,will/would + 動詞 — 條件及其結果 | if I were you, I would think |
| 現在完成時 have/has + 過去分詞 — 過去發生但與現在仍有關聯的事 | have been · have you been · has done |
| 被動語態 be + 過去分詞 — 強調發生了什麼,而不是誰做的 | being asked · be featured · were caused |
需要注意的發音
說話人用了 151 次縮略和弱讀形式,例如 you're, don't, we're。請按聽到的簡短形式來說。
- “sh” 和 “zh” 音: migration /maɪˈɡɹeɪʃ(ə)n/, implementation /ˌɪmplɪmənˈteɪʃən/, ownership /ˈoʊnɚʃɪp/, intention /ɪnˈtɛnʃn̩/, relation /ɹɪˈleɪʃ(ə)n/
- 長單詞——注意重音位置: disagreement /dɪsəˈɡɹiːmənt/, complicated /ˈkɑm.plɪˌkeɪ.tɪd/, complexity /kəmˈplɛk.sɪ.ti/, communicate /kəˈmjuːnɪkeɪt/, experiment /ɪkˈspɛɹ.ə.mənt/
中文母語者容易讀錯的音:
- /v/ — 不要讀成 /w/,上齒輕觸下唇: interviewer /ˈɪntɚvjuɚ/, convince /kənˈvɪns/, solve /sɒlv/, reverse /ɹɪˈvɝs/, behavioral /bɪˈheɪvjəɹəl/
- 詞尾子音 — 要讀清楚,後面不要加母音: component /kəmˈpoʊ.nənt/, disagreement /dɪsəˈɡɹiːmənt/, convince /kənˈvɪns/, scope /skəʊp/, implement /ˈɪmpləmənt/
- /r/ 和 /l/ 的區別: behavioral /bɪˈheɪvjəɹəl/, relative /ˈɹɛl.ə.tɪv/, strongly /ˈstɹɔŋli/, principle /ˈpɹɪn.sɪ.pəl/, relation /ɹɪˈleɪʃ(ə)n/
如何用這段影片練習
- 先完整聽一遍影片,不要開口,記下不認識的單詞。
- 先用 0.75 倍速逐句跟讀,熟練之後再回到正常速度。
- 錄下自己的聲音並與原聲對比,特別注意 disagree, deadline, component 這類單詞。
什麼是跟讀法?
跟讀法 (Shadowing) 是一種有科學依據的語言學習技巧,最初開發用於專業口譯員的培訓,並由多語言者Alexander Arguelles博士普及。這個方法簡單而強大:您在聽英語母語原聲的同時立即大聲重複——就像是一個延遲1-2秒緊跟說話者的影子。與被動聽力或語法練習不同,跟讀法強迫您的大腦和口腔肌肉同時處理並模模擬實的講話模式。研究表明它能顯著提高發音準確性,語調,節奏,連讀,聽力理解和口語流利度——使其成為雅思口語備考和真實英語交流最有效的方法之一。












