Technical leadership
A senior platform engineer interview on leading a platform as a product for internal teams, from earning adoption and retiring old systems to outages, guardrails, and exceptions.
These prompts are original practice. They are not questions from a named employer or a live hiring loop.
1.You lead a platform team at an insurance company. Your new deploy pipeline is better than the old one on paper, but after a quarter only three of twenty-five teams have moved. What do you do first?
2.Your platform team at a travel site measures itself by tickets closed each week. Teams still say the platform slows them down. As the staff engineer, which change to the team's goals do you propose?
3.A product team at a freight company built its own deploy scripts instead of using your platform. One of your engineers wants to report them to their director for leaving the paved road. What do you suggest?
4.You are drafting the first yearly roadmap for the internal developer platform at a regional bank. What should it be grounded in?
5.Your platform team at a clothing retailer has no product manager. Requests arrive in chat, email, and hallway asks, and teams feel ignored. As the lead, what is a good first step?
6.A platform team at a video streaming company plans to rebuild its internal developer portal. You are the staff engineer. Before any design work starts, what do you push for?
7.Teams at a health records company rate your platform's documentation poorly. Your engineers say every feature is documented. What do you do?
8.A new engineer on your platform team at a game studio asks why the team spends time interviewing product engineers. They say the platform team already knows what users need. How do you respond?
9.Your platform team at a pharmacy chain ships a new feature. Most teams who try it hit a confusing error in their first five minutes, and few come back. What do you prioritize next?
10.Your team at an online marketplace built a new secrets service. Fewer than ten percent of teams use it. Your director offers to make it mandatory next month. What do you say?
11.You want teams at a payments company to adopt a new service template. Which team is the best first adopter?
12.Two respected engineers at a media company openly doubt your new build system in a public channel. Several teams are waiting to see what they think. What do you do?
13.At a ride-hailing company, three past migration guides for platform libraries were mostly ignored. One of your engineers proposes a guide for the new logging library. What do you suggest instead?
14.You run the platform team at a food delivery firm. A new caching feature is ready, but product teams are deep in a major launch for the next six weeks. What is the best plan for adoption?
15.A director at a gaming company asks why adoption of your platform is slow. Your honest view is that the platform still lacks support for scheduled jobs, which many teams need. What do you tell them?
16.You are the staff engineer on a platform team at a home goods retailer. An early adopter team is unhappy because a feature you promised them slipped by a month. What do you do?
17.You want every team at a hospital network to clean up noisy alerts. You have no authority over those teams. What approach works best?
18.A platform engineer at an energy company proposes awards for teams that adopt the new pipeline fastest. Adoption is stalled because the pipeline cannot run database migrations. What do you advise?
19.You lead the retirement of an old deploy system at an airline. Forty teams still use it, and some built custom scripts on top of it. What should come first?
20.Halfway through retiring an old deploy system at a telecom, you find a payroll export job that runs through it. No team claims the job. What do you do?
21.Eighty-five percent of teams at a bank have left the old deploy system. The last few teams have tight deadlines and fragile services. What is the best approach for the tail?
22.The old deploy system at an insurance firm has manual approval gates that two teams use for regulated releases. The new system has no such feature. What do you do?
23.Your plan to retire an old deploy tool at a hospital network gave teams four months. One month in, half the teams have not started. What do you do?
24.A team at a shipping company asks for a one-year delay on leaving the old deploy system because of a large product launch in three months. What is the best response?
25.You are retiring an old deploy system at a water utility. Usage logs show almost no traffic, but you are not sure every user is known. Which shutdown plan is safest?
26.Leadership at a consumer bank asks why the deploy system migration is two quarters late. The main cause is that you underestimated custom integrations. What do you tell them?
27.An engineer on your team at a ticketing company wants to rewrite the new deploy system's internals before finishing the move off the old system. The current internals work fine; the goal is cleaner code. What do you advise?
28.You are leading the move off an old deploy tool at a news site. A team says their service will break on the new tool but cannot say why. What do you do?
29.You are leading the move from an old continuous integration (CI) system to a new one at an outdoor gear retailer. A team asks to move after the deadline because their pipeline has 400 custom steps. What do you do first?
30.The old deploy system at a cable company has finally been shut down. What is the most useful thing for you, as the lead, to do next?
31.You are retiring an old deploy system at a mortgage lender. A team's custom rollback script only runs on the old system. They say this blocks their move. What do you do?
32.Your platform team at a payments company has three big asks: GPU nodes for one team, faster builds for five teams, and image signing that security needs to pass a required audit in ten weeks. You can do one this quarter. What do you do?
33.A product director at an insurance company asks your platform team to build a feature by next month. It does not fit this quarter's plan. How do you say no well?
34.Your platform team at a fintech spends half its time on manual requests like creating databases. Leadership wants more new features from the team. What do you propose?
35.Two product teams at a travel company want conflicting changes to the shared ingress setup your team owns. Both say theirs is urgent. What do you do?
36.Your platform team at a game studio spent last quarter shipping dozens of small requests and no major project. Teams still complain about the same big problems. What change do you make?
37.A senior engineer at a hospital network asks your team to support a new database engine. Only their team wants it, and supporting it would take a full engineer. What do you tell them?
38.You lead a platform team at a grocery chain. Midway through the quarter, your manager adds a new top priority without removing anything. What do you do?
39.Your platform team at an online school promised a self-service database tool by quarter end. Six weeks out, the full scope clearly will not fit. What do you do?
40.You are planning a year-long effort at a bank to move teams onto a new compute platform. How do you sequence the work?
41.An executive at a media company asks your platform team to build an internal dashboard they saw at a conference. It would take a month. What do you do?
42.Your manager at a logistics company tells a vice president your team will finish the cluster upgrade by March. You were not asked, and you think March is not possible. What do you do?
43.Your shared deploy system at a sporting goods retailer fails, and no team can deploy for three hours. Once it is fixed, what do you prioritize as the senior platform engineer?
44.During a platform-wide outage at a bank, product teams flood your channel with questions. Your engineers are busy debugging. What do you do as the lead?
45.In the review after a platform outage at a media firm, a product team lead says the platform team is careless. How do you respond?
46.A junior engineer's config change at a payments company caused a cluster-wide outage. They are very upset. What do you do as the staff engineer?
47.After a platform outage at an airline, the review listed twelve action items. Three months later, two are done. What do you do?
48.Your platform team at a shipping company caused two outages in a month. Your director wants to freeze all platform changes until next quarter. What do you propose?
49.A platform outage at a hospital network was triggered by a cloud provider failure in one region. What do you put in your summary for product teams?
50.You are incident lead for a platform outage at a department store chain. Customers cannot check out. One engineer wants to keep digging for the root cause; another has a tested rollback ready. What do you decide?
51.Teams at an insurance company lost trust in the platform after a bad outage. Some now talk about running their own clusters. As the staff engineer, what do you do?
52.During a platform outage at a gaming company, responders found that the runbook for the failing component was out of date. What should come out of the review?
53.A shared DNS change by your team at a telecom broke internal service lookups for an hour. Product teams were not told until forty minutes in. Besides the technical fix, which follow-up matters most?
54.You are the senior platform engineer during a large outage at a credit union. You know the failing system well and want to start debugging. When is that the wrong move?
55.Your platform team at a regional bank is designing a golden path for new services. A teammate wants it to cover all six languages the company uses from the start. Most new services use one language. What do you suggest?
56.A security lead at a fintech wants your platform to block every deploy with any known vulnerability. Many are low risk and have no fix yet. What do you propose?
57.Most teams at a media company use your golden path for services. A few teams found it too rigid and left it. What do you do?
58.You plan a guardrail at a travel company that rejects deploys without health checks. About thirty percent of services lack them today. How do you roll it out?
59.A staff engineer on another team at an e-commerce company says your deploy guardrails add twenty minutes to every release. You think the guardrails matter. What do you do?
60.Your platform team at a school software company keeps adding rules to the golden path. Teams now say it is hard to understand. What do you do?
61.Your golden path at a bank sets a database size by default that costs far more than most teams need. A team lead points this out. What do you do?
62.Your team at a hotel booking company built a guardrail that blocks public storage buckets. A week later, a team says it blocked a bucket that serves their public website. What do you do?
63.A team at an insurance company asks for an exception from the platform's standard logging. Their service handles very high-volume event data. What do you do first?
64.You grant a team at a sporting goods retailer an exception to the platform's network rules. What should come with it?
65.Five teams at a gaming company now have exceptions to the same platform rule. What does this most likely show?
66.A team at a payments company asks to skip the platform's security scan so they can ship before a deadline next week. The scan has never failed for them. What do you do?
67.A well-liked senior engineer at a media firm wants an exception from the golden path's deploy process. Their only reason is that they prefer their own scripts. What do you do?
68.A team at a bank has had a platform exception for two years. The original reason no longer applies. What do you do?
69.A team at a logistics company wants an exception to run without autoscaling limits on a node pool used only by their service. You think it is risky for them. Their director supports it. What do you do?
70.A team at a fintech asks for an exception to the platform's Friday deploy block. They must ship a regulatory fix by Friday night. What do you do?
71.An exception request at a car rental company sits in your team's queue for three weeks while the requesting team waits. What do you change?
72.You are writing a design doc at a hospital network for a new shared service mesh. Who should review it before the decision is made?
73.In a design review at a fintech, a senior engineer strongly disagrees with your choice of build tool. You have heard their points and still think your choice is right. You own the decision. What do you do?
74.Your design doc at an e-commerce firm proposes a new config system. It is thirty pages long, and nobody has commented in two weeks. What do you do?
75.In a design review at a gaming company, a junior engineer raises a concern about failure handling that others brush aside. You think the concern is valid. What do you do?
76.Your team at a bank must choose between two queueing systems. Both would work, and the team has argued for three weeks. You are the lead. What do you do?
77.You are reviewing a peer's design doc at a media company for a large platform change. You see a serious flaw in how it handles partial failures. What do you do?
78.You run a design review at an airline. One engineer dominates the discussion, and others say almost nothing. What do you do?
79.You are proposing a new artifact storage design at a freight company. What belongs near the top of the design doc?
80.Your platform team at a travel company must choose a new container registry. You lack data on future load. What is the best approach?
81.A year ago at a fintech, your team chose a service discovery tool. New facts show it no longer fits. A teammate says switching would mean admitting a mistake. What do you say?
82.Your team at a hospital network makes most architecture decisions in chat threads. New engineers cannot tell why things are built the way they are. What do you introduce?
83.You must decide at a department store chain whether to switch all services to a new deploy strategy two weeks before the holiday code freeze. Early test data is mixed. What do you do?
84.Your team at a bank is choosing an approach for a new secrets rotation system. One approach is easy to undo later; the other locks in a storage format. How should that shape how you decide?
85.You chose a new build cache design at a logistics company six months ago. You set a review date then. The date has come and nobody is complaining. What do you do?
86.Your platform at a media company runs on a cluster tool version that loses vendor support next year. Leadership sees no reason to spend time on it. How do you make the case?
87.Your team at an insurance company wants a quarter to clean up the platform's code. Product leaders want features. What do you propose?
88.Your platform at a gaming company has many small pieces of technical debt. Which should you fix first?
89.A build tool your platform maintains at a bank has grown so complex that only two engineers can change it. Leadership calls it working. How do you frame the problem?
90.Your platform project at a bank is three weeks behind, with launch in eight weeks. Stakeholders do not know yet. What do you do first?
91.Your project at a hospital network is slipping because another team has not delivered an API you depend on. What do you do?
92.You lead a migration at an e-commerce company. Last week's status was green, but you now know two key pieces are at risk. What do you report this week?
93.A project you lead at a media company has slipped twice. Stakeholders are losing patience. What do you change?
94.Two weeks before a cluster upgrade at a cable company, testing shows a networking plugin breaks on the new version. Product teams have planned around the date. What do you do?
95.You are the staff engineer on a platform team at a fintech. Every design question goes through you, and work waits days on your reviews. What do you do?
96.A mid-level engineer at a bank wants to lead the next platform project. You think they are ready but may make some mistakes. What do you do?
97.An engineer you mentor at a home goods retailer brings you a design problem. You already know a good answer, and it is not urgent. What do you do?
98.A new engineer on your platform team at a game studio is struggling on their first on-call shifts. What do you do?
99.Your platform team at an insurance company relies on one engineer who knows the deploy system best. They are burned out. What do you do?
100.You are a staff engineer at a media company. Your team keeps asking you to write all the hardest code. What do you do?
101.Your platform team at a fintech has a rule that only you can approve production config changes. Changes now wait a day or more on you. What do you do?
102.You are leaving a platform team at an insurance company for another team in a month. What matters most before you go?
103.You delegated ownership of the platform's cost reporting at a car rental company to a senior engineer. Their first design differs from what you would build, but it meets the needs. What do you do?
104.Your director at a fintech wants the platform team to build its own service mesh. You think an existing open-source option fits better. What do you do?
105.A principal engineer at an e-commerce firm proposes a platform design you believe will cause outages. They are more senior than you. What do you do?
106.Your platform team and the security team at a bank disagree on how strict image scanning should be. Each side has real data behind its view. What do you do?
107.Code reviews on your platform team at a ticketing company often catch the same config mistakes. Reviewers are tired of pointing them out. What do you do?
108.A product team at a telecom asks your platform team to own their service's on-call because it runs on your platform. What do you say?