In this article
Before you start
This page sits in the engineering manager interview questions guide, and it assumes you have met two ideas.
In this page, "top talent" means the strongest performers on a team, and the interviewer is usually an engineering manager (EM) or a director. The question is reported by candidates in people-management rounds, and the people management round explains how that round works.
Why interviewers ask this
The question separates managers who grow people from managers who hand out titles, and getting it wrong loses the strongest people on the team. Google's Project Oxygen analysed over 10,000 data points about its managers from performance reviews and feedback surveys, plus hundreds of pages of interview notes. It found that teams with highly effective managers were more likely to achieve better results, be happier and have lower turnover. Two of its listed manager behaviours are "Be a good coach" and "Support career development and discuss performance."
A weak answer says "I give them the hardest tickets and push for promotion." A strong answer says the manager first finds out what each person wants, because high performers differ, and then builds a plan around that. A Staff-level answer also names the cost of the plan: the work has to come from somewhere, and the strongest people are the easiest to overload.
Mentoring top talent from first principles
- Growth trajectory
- A growth trajectory is the pace at which a person wants their scope and impact to increase in a given phase of their career, either steep, with constant new challenges, or gradual, with deepening mastery and stability. It is not the same as promotion, and a person can change trajectory over time.
- Stretch assignment
- A stretch assignment is a piece of work just beyond a person's current skill, given with a clear learning goal and support, so that doing the work builds a new capability. It differs from extra workload because it comes with mentoring and review.
- Sponsor
- A sponsor is a senior person who uses their own standing to put a junior colleague forward for opportunities, such as a project, a promotion discussion or a visible meeting. It differs from a mentor, who advises but does not necessarily advocate for the person.
In plain words
Mentoring top talent means helping your strongest people grow in the direction they choose, with support from you as their manager. Think of a coach with two runners, a sprinter and a marathoner. Both are fast, and both need a coach, but the training plans differ. The analogy breaks in one place. People are not fixed types. The same engineer can be a sprinter this year and a marathoner after a new baby or a move.
In precise terms, mentoring top talent starts with learning each strong performer's growth trajectory. Then you supply the stretch work, feedback and sponsorship that match it.
Step-by-step description
- Career conversations show what the person wants.
- The plan fits the growth trajectory.
- Stretch work comes with mentoring and review.
- Sponsor the person, then revisit the plan yearly.
- Career conversations are three structured talks that tell you what a person wants before you plan anything.
- Stretch work is the main lever, and it works best when combined with mentoring and review.
Step 1: Hold career conversations to learn what each person wants
Career conversations are three structured talks that tell you what a person wants before you plan anything. Radical Candor describes them as a Life Story conversation about what motivates and drives the person, a Dreams conversation about where they want to be at the pinnacle of their career, and a Career Action Plan for the present. It suggests spacing them a week or two apart and revisiting them annually, since goals change.
Input: a person you manage. Output: a written picture of what they value, where they hope to end up and what they want in the next year. Your regular one-on-ones are where the talks happen, and they only work if the person trusts you enough to be honest. You now know what the person wants. The next step turns that into the right plan.
Step 2: Match the plan to the growth trajectory, not to a promotion
The plan for a top performer depends on the trajectory they want, and a promotion is only one possible path. Kim Scott of Radical Candor describes people in superstar mode as needing to be challenged and given new opportunities to grow constantly, while people in rock star mode do not want the next job if it takes them away from their craft. She warns that promoting someone into a role they do not want or are not suited for means you lose them, or end up firing them.
Steep growth is also not the same as management. Plenty of individual contributors stay on a steep trajectory their whole careers, and it is about increased impact over time. Scott's reframe is to stop asking whether a person has high or low potential and ask what growth trajectory each person wants to be on right now.
| Mode | What the person wants | What you give | How it goes wrong |
|---|---|---|---|
| Superstar mode (steep) | New challenges and wider scope | Stretch work, new teams, sponsorship | Boredom, so they leave |
| Rock star mode (gradual) | Mastery and stability | Hard technical ownership, recognition | A promotion they hate |
Step 3: Give stretch work, and pair it with mentoring and review
Stretch work is the main lever, and it works best when combined with mentoring and review. The University of Maryland's Smith School summarises research that on-the-job experience may be the primary vehicle for learning critical leadership skills, and advises using challenging experiences in combination with active mentoring and review processes.
The 70-20-10 model is a heuristic about where learning tends to come from: about 70% from job experience, 20% from interactions with others and 10% from formal courses (Training Industry). It was created in the 1980s by researchers working with the Center for Creative Leadership. It describes where managers reported their learning came from, so it is not a budget for a mentoring plan. Clardy (2018) reviewed it under the title "a fact in search of evidence".
What the heuristic does support is the shape of a plan: the stretch work is the centre, and mentoring and courses support it. For two imagined engineers the content differs even though the shape is the same. The superstar-mode engineer leads a cross-team project with a sponsor outside the team. The rock-star-mode engineer owns the hardest system in the area for the year, with a peer review of design work.
Stretch work has a cost. The same Maryland page notes that companies may not want to move a person who is doing a fine job in a mission-critical position just to give a rising colleague room to grow. Plan the handover before you promise the work.
Step 4: Sponsor the person, then revisit the plan
A mentor advises, and a sponsor acts, so top talent usually needs both. Sylvia Ann Hewlett, in a Harvard Business Review (HBR) On Leadership podcast episode (October 2023), calls mentorship a one-way gift of guidance and advice, and says sponsorship needs to be a reciprocal relationship. Your job as manager is to put your strongest people forward for visible work and to find sponsors beyond you when your own reach ends.
Then you revisit the plan. Scott notes that most people shift between steep and gradual trajectories in different phases of their lives, so a label you gave someone last year may be wrong now. The yearly revisit of the career conversations is how you catch that. Treat each person's mode as a current state, not a permanent label.
Knowing it is half the job. Saying it is the other half.
Interview Clozo is an AI interview copilot for real interview rounds: it hears the question and puts a strong answer on your screen, and you say it in your own words. See the features or how it works.
Start Free Trial →The Staff-level answer
How do you mentor top talent?
I start by asking each person what growth they want, because top performers differ. Some want wider scope and constant new challenges. Others want to deepen their craft and stay where they are. For the first group I give stretch work, like leading a cross-team project, with a mentor and regular review, and I sponsor them for bigger scope. For the second I give hard technical ownership and recognition, not a management title. I hold the career conversations a week or two apart, then revisit them every year, because goals change. The risk is overloading my best people, so I give each piece of work a clear learning goal and check it against their capacity.
What makes this Staff-level is that it refuses to assume promotion is the goal, names a plan for each type, and ends on the risk instead of the plan. Anchor it in one real example with its outcome, because an interviewer's first follow-up is "tell me about a time".
Text description of the diagram
- A strong answer states the principle, then one real example with its outcome.
- It ends on the cost or risk.
Illustrative example: a weak and a strong answer
This is an illustrative example, a composite and not a real candidate. The weak candidate says that top performers get the hardest tickets, a fast promotion path and regular praise. There is no person, no outcome and no cost in it, so the interviewer cannot tell whether it ever happened.
The strong candidate states the principle first: ask what each person wants. Then the candidate tells one story in plain steps. An engineer who was the best on a team of six said she wanted to stay hands-on. The manager gave her ownership of the hardest ingestion service for two quarters, paired her with a staff engineer as mentor, and held a weekly design review. Illustratively, pages per week fell from about twelve to three, and she kept her seat on the team and made a promotion case on the technical track. The candidate ends on the cost: the work had to be taken off another engineer first.
The words that make the difference are the person's own goal, the specific work, the support around it, a measurable result and the stated cost.
The rubric
| Dimension | Weak answer | Strong answer | Staff answer |
|---|---|---|---|
| Correctness | Give them harder work and a title | Ask what they want, then plan | Match plan to trajectory and revisit yearly |
| Depth | One generic plan for everyone | Different plans for different goals | Names stretch, feedback and sponsorship roles |
| Trade-offs | None mentioned | Notes overload risk | Plans handover and capacity before offering work |
| Communication | Vague and abstract | Gives one concrete example | Gives two contrasting examples in plain words |
| Follow-up readiness | Stuck when asked about people who want to stay | Handles the stay-in-craft case | Also handles a changed goal and a departure |
Three variants of this question
Your best engineer wants to stay an engineer. How do you keep growing them?
You give them deeper technical ownership, recognition and harder problems, and you make clear that staying in the craft is a respected path. Promoting them into a role they do not want risks losing them. The tech lead interview questions cover the individual-contributor growth path in more depth.
How do you mentor a high performer who wants a promotion before they are ready?
You show the gap with specific examples and give a stretch assignment aimed at it, with review, so the person can build the missing skill. Then you agree what evidence would change the answer.
How do you develop and retain high performers without neglecting the rest of the team?
You keep the same career conversations with everyone, because the question applies to every trajectory, and you spread stretch work and sponsorship instead of concentrating it. Retention follows from the plan fitting what each person wants. The growing engineers guide covers the wider team.
Common mistakes
- "I promote my best people as fast as I can." (wrong) Some of them do not want the next job. (right) "I ask what growth each person wants, and promotion is one option."
- "Stretch work means giving them more tickets." (wrong) Extra load is not development. (right) "Each piece of stretch work has a learning goal, a mentor and a review."
- "My top performers are high potential and the rest are not." (wrong) That label has no place for people who are great and want to stay. (right) "I ask what trajectory each person wants right now."
- "Mentoring and sponsoring are the same thing." (wrong) A mentor advises, a sponsor advocates. (right) "I mentor them, and I put them forward for the visible work."
- "Top people will stay on their own." (wrong) Google's data ties lower turnover to highly effective managers. (right) "I keep the career conversation going every year."
Follow-up questions
How do you tell whether someone is in superstar or rock star mode?
You ask, using the career conversations, instead of guessing from their output. Treat the answer as current, because the same person can change mode as life changes.
What is the difference between a mentor and a sponsor?
A mentor gives advice and guidance, usually one way. A sponsor uses their own standing to put the person forward, and the relationship works best when both sides gain.
Is the 70-20-10 model proven?
No. It was created in the 1980s by researchers working with the Center for Creative Leadership (Training Industry), and Clardy (2018) reviewed it under the title "a fact in search of evidence". Use it as a heuristic about where learning tends to come from, not as a budget.
How often should you hold career conversations?
Hold the three talks a week or two apart, then revisit them annually, as Radical Candor suggests. Short check-ins in your regular one-on-ones keep the plan alive in between.
What if a top performer leaves anyway?
Hold an exit conversation and ask what you could have learned: did the plan match their trajectory, and was the stretch work real? Arrange a knowledge handover so the team does not lose context, and keep the relationship, since people sometimes return or refer others. Then check for a pattern across regretted departures, such as the same team or the same missing growth path. The low performer guide covers the opposite problem, and the team health measures show early signs of disengagement. Your wider stance is the subject of what's your management philosophy.
- Top talent is not one type: ask each person what growth trajectory they want now.
- Steep growth is not the same as promotion, and promoting someone into a role they do not want can lose them.
- Pair stretch work with mentoring and review, with a clear learning goal.
- Sponsor your strongest people, and revisit the plan every year.
- Anchor your answer in one real example with its outcome and its cost.
Sources
- Following the data: The research behind great managers — Google re:Work, 1 Aug 2025
- Balancing Growth and Stability: Why Your Team Needs People On Both Steep and Gradual Growth Trajectories — Radical Candor (Kim Scott), 9 Jan 2023
- 6 Ways to Roll Out Radical Candor Like a Boss (Career Conversations) — Radical Candor, 26 Oct 2026
- The 70-20-10 Model for Learning and Development — Training Industry, Inc., 26 Oct 2026
- Developing managerial talent through stretch assignments — Robert H. Smith School of Business, University of Maryland, 26 Oct 2026
- The Benefits of Career Sponsorship Go Both Ways — Harvard Business Review, HBR On Leadership, 18 Oct 2023
- 70-20-10 and the Dominance of Informal Learning: A Fact in Search of Evidence — Human Resource Development Review 17(2) (Clardy), 1 Jun 2018
