Sample run
See it on a sample CV
One CV, the interview questions it gets you asked, an answer with its score and one fix, and the second try.
From the CV
Mentored two junior engineers; introduced a design-review template the team now uses for every new service.
The question
You introduced a design-review template that the team now uses for every new service. What problem prompted it, and how did you get a 40-person engineering team to adopt it?
One fix later
5.5 → 8.1
Target met: one clear main point, stated first
Sample run on a fictional CV, through the app's own question generator and scoring. Not a live demo. The answers are read by an AI voice from a written script; the app's pipeline transcribed and scored the recordings.
The CV
Priya Sharma, backend engineer: a made-up person, and a CV written for this run.
The questions it asked
From the question bank the app wrote for this CV. Each one comes from a line of it.
Batch to event-driven invoicing
Walk me through how you led the move of invoice generation from a nightly batch job to an event-driven service on Kafka. How did you plan the cutover and what did you own?
From the CV: Led the move of invoice generation from a nightly batch job to an event-driven service on Kafka; invoices now go out in under two minutes instead of the next morning.
Duplicate-charge post-mortem
You ran the post-mortem after a duplicate-charge incident affecting 312 customers. What was the root cause, and why did you choose idempotency keys across all payment endpoints as the fix?
From the CV: Ran the post-mortem after a duplicate-charge incident that affected 312 customers; added idempotency keys across all payment endpoints.
This run's question
Answered in steps 3 and 4Design-review template and mentoring
You introduced a design-review template that the team now uses for every new service. What problem prompted it, and how did you get a 40-person engineering team to adopt it?
From the CV
Mentored two junior engineers; introduced a design-review template the team now uses for every new service.
Settlement service at 1.2M a day
Tell me about building the merchant settlement service in Go that handled about 1.2 million transactions a day. What were the main design concerns at that volume?
From the CV: Built the merchant settlement service in Go, handling about 1.2 million transactions a day.
Scaling on queue depth
In your open source autoscaler you scaled workers on message-queue depth instead of CPU. Why was that the better signal, and where might it fall short?
From the CV: A small Kubernetes controller that scales workers on message-queue depth instead of CPU; about 300 GitHub stars.
The first answer
Answered out loud. Then the score, what went well, and the one fix to make.
Interview question
You introduced a design-review template that the team now uses for every new service. What problem prompted it, and how did you get a 40-person engineering team to adopt it?
First answer · 0:53 Transcript
So, um, okay, this was at Northwind. We had a lot of teams and like a lot of services and it was a busy time. We were hiring a lot also. And um, the thing is people would just start building things and then later someone would ask questions in code review like about Kafka or databases or whatever and it became a whole discussion. I was also mentoring two juniors around then, which was actually really nice. They were both quite sharp and. And they also didn't always know what to think about first. So, uh, I made a kind of template, a doc. It had some sections, the problem options, that kind of thing. And uh, I put it out there and over time people started using it I think because it was useful and my manager was also supportive. So yeah, it kind of became the standard. So that's uh, basically the story.
Overall score
Overall score 5.5 out of 10
Almost there
- Clarity
- 4.9/10
- Structure
- 5.0/10
- Delivery
- 7.6/10
177 wpm · 5 fillers · 0 pauses mid-sentence
The answer gestures at both parts of the question but never gets concrete about the problem, the template's contents, or the actual adoption tactics with a 40-person team.
Strength: The chronological setup (busy hiring period, recurring code-review friction, template created in response) is easy to follow up to that point.
The one fix
Clarity — One clear main point, stated first
The second try
The same question again, working on that one fix.
Working on: Clarity — One clear main point, stated first
Second try · 0:51 Transcript
I got a team of 40 engineers to adopt it by making it the easiest way to get a design approved, not by asking people to use it. The problem was rework teams would build a service, and only in code review would someone ask why it needed a new database. So work got redone. I wrote a one page template. The problem, the options we rejected, the data, and the rollout plan. Then I did three things. I used it on my own next service. So there was a real example. I walked through it at our engineering all hands, um, in about 10 minutes. And I asked two senior engineers to review designs only when they came in that format. Within a quarter, every new service was using it. And those design questions moved out of code review to before anyone wrote code.
Target met
Clarity — One clear main point, stated first
You now open with a single clear headline claim about how adoption happened, and the previous feedback about naming concrete adoption steps (own use, all-hands walkthrough, senior reviewer gatekeeping) and a specific problem has been directly acted on.
Opening sentence: 'I got a team of 40 engineers to adopt it by making it the easiest way to get a design approved, not by asking people to use it' states the core claim immediately, and the rest of the answer (the problem, the template, the three actions) supports that claim without introducing a competing one.
Clarity score
4.9 → 8.2 (up)
The drill landed, and your broader Clarity score rose too.
Next: Add one named concrete incident (specific service, specific rework cost) to make the problem statement as vivid as the adoption steps.
Overall: 8.1
