"Tell me about a time you received constructive feedback."
A simple question. A Staff-level candidate. Should be easy.
"I led development of a new service at Amazon. Hard deadlines, exciting technical challenges... I realized that documentation and handoff are full requirements..."
The candidate was rejected on the behavioral loop.
Why? Because at Senior+ levels, story choice matters more than story structure.
Why you should listen to me
My name is Gilad. I was an engineering leader at Amazon and Meta... After hundreds of sessions, my intuition told me that delivery matters — but not as much as everyone fears. The main reason for rejection is the wrong story, not weak delivery.
The STAR/CARL problem at Senior+ levels
There is an interesting problem that gets worse as your level rises. The more successful your career, the larger this problem becomes.
When you aim for a Staff+ role, you are not just looking for a job. You are targeting work at a specific level.
Say you are interviewing for Senior Manager. You have probably coached a "low performer" on your team more than once. Which story do you pick? The secret is that your story must show the difference between you and a typical line manager.
Or imagine you are targeting Staff Engineer. You talk about coaching a junior on pull-request hygiene. The interviewer knows most Senior engineers do that every week. Instead, show how you coached a new TL and helped them become a leader.
A 5-step prep algorithm (Pre-compute)
Over the years I built a 5-step algorithm that helps clients prep for interviews. Here is how you can use it for your upcoming loop.
Step 1: Brainstorm stories (mapped to criteria)
Start by understanding what the company values. Study the criteria: ambiguity, conflict management, leading cross-team initiatives, technical trade-offs at scale, stakeholder management, dealing with failure.
Open a doc and dump everything that comes to mind. Aim for 20–30 stories.
Step 2: Define scope and difficulty
For each story, pin down what was hard. Be specific: "a 3-week deadline for a project that normally takes 3 months, during a Q4 code freeze" shows real difficulty.
Step 3: Gather evidence
Most candidates skip this step. You need facts — quantitative, qualitative, and organizational evidence.
- Number of teams involved (7 teams across 3 orgs)
- Budget or revenue impact ($2.3M saved)
- Stakeholder level (VP approval required)
- Recognition received (called out at all-hands by the CEO)
- First-of-its-kind experience at the company
- Policy or process change
Go back to old emails, messages, and reviews. Find the receipts.
Step 4: Hard filtering
Any story without both clear difficulty AND convincing evidence at the target level — delete it. You should end with 8–10 stories.
Step 5: Craft a headline
Every story needs a compelling headline. Instead of: "Let me tell you about a time I resolved conflict" → "Let me tell you how two VPs wanted opposite architectures, and I had 3 weeks to align them."
Before and after examples
Let's look at examples of applying the 5-step algorithm.
Meta — Conflict resolution (Staff level)
Rejected stories
- Disagreed with a PM on priorities — routine discussion; the PM agreed quickly.
- Architect wanted microservices; I wanted a monolith — a tech debate inside one team.
- Security blocked a launch — standard process.
- Manager assigned conflicting priorities — the manager's problem.
The story that passed
Two teams wanted different API designs.
- Difficulty: 12 engineers, both APIs valid, 6-month deadline at risk, no decision-maker.
- Evidence: $3M budget at risk, 15 downstream services, escalated to director, solution became a template.
Amazon — "Have Backbone; Disagree and Commit" (Staff level)
Rejected stories
- Held my ground with a director — the director changed course after one meeting.
- Argued build vs buy — we voted; I agreed.
- Challenged the roadmap — healthy debate; I dropped it.
- Challenged the on-call schedule model — process tweak for 8 people.
The story that passed
I pushed back against shutting down a project I built.
- Difficulty: director wanted to kill a 2-year project, declining metrics, future potential, team morale.
- Evidence: $1.2M invested, asked for 30 days, lost on the data, helped wind down in 2 weeks instead of 2 months.
Answers to your questions
What if I don't have enough evidence?
Dig deeper (email, archives), start collecting now, or admit the gap — you may not yet have target-level experience.
How specific should evidence be?
Very specific. "A large team" → "27 engineers across 4 teams." "A hard deadline" → "6 weeks during a Q4 code freeze."
Can one story cover multiple criteria?
Yes, if you have different evidence angles. A hard migration can show leadership, ambiguity, and technical excellence.
What if my evidence is confidential?
Anonymize: "We saved a 7-figure amount" or "Cut costs by 70%." Magnitude matters more than details.
How many stories do I really need?
6–10 stories with 3–5 pieces of evidence each. Fewer than that means you have a gap.
Does this work for non-technical roles?
Absolutely. Only the evidence changes: for PMs — users and revenue; for designers — A/B tests; for sales — deal size.
What's next?
You have the algorithm. Now use it.
Block three hours this week. Open a doc. Walk all five steps. Pay special attention to Step 3 — Evidence. Find those receipts.
You will be surprised how many of your "great stories" lack convincing evidence. That is good. Better to learn it now.