How to Find a Mentor in Tech

Advice on how to find a mentor in tech often starts with "reach out to senior people on LinkedIn". Most of those messages go unanswered, because they ask a stranger for an open-ended commitment. Mentoring works better when you understand what it really gives you and ask for something small and specific. This guide starts with the evidence: two large reviews of mentoring research, and answers from about 24,000 tech workers in the 2025 Stack Overflow survey (Stack Overflow survey 2025).

import pandas as pd

so = pd.read_csv("results.csv", low_memory=False)          # 2025 Stack Overflow Developer Survey
# Job satisfaction factors were ranked 1 (most important) to 15; _4 = expert mentors, _5 = mentoring juniors
ranked = so.dropna(subset=[f"JobSatPoints_{i}" for i in (1, 2, 3, 4, 5)]).copy()
ranked["experience"] = pd.cut(ranked.WorkExp, [0, 2, 5, 10, 20, 100], labels=["1-2 years", "3-5", "6-10", "11-20", "21+"])

print("M1 how highly people rank mentors among 14 job satisfaction factors, by years of experience")
print(f"   {'experience':<11}{'mentors in top 3':>18}{'average rank':>14}{'mentoring others in top 3':>28}{'n':>8}")
for band, g in ranked.groupby("experience", observed=True):
    print(f"   {band:<11}{(g.JobSatPoints_4 <= 3).mean():>18.0%}{g.JobSatPoints_4.mean():>14.1f}"
          f"{(g.JobSatPoints_5 <= 3).mean():>28.0%}{len(g):>8,}")

print("\nM2 learned to code in the past year from a colleague or on-the-job training")
learn = so.dropna(subset=["LearnCode"]).copy()
learn["from_colleague"] = learn.LearnCode.str.contains("Colleague or on-the-job training", regex=False)
learn["arrangement"] = learn.RemoteWork.map(lambda v: "fully remote" if v == "Remote" else "in-person" if v == "In-person"
                                            else "hybrid or flexible" if isinstance(v, str) else None)
for arrangement, g in learn.dropna(subset=["arrangement"]).groupby("arrangement"):
    print(f"   {arrangement:<20}{g.from_colleague.mean():>6.0%}   n={len(g):,}")

print("\nM3 people managers among respondents, by years of experience (who could mentor you)")
icpm = so.dropna(subset=["ICorPM", "WorkExp"]).copy()
icpm["experience"] = pd.cut(icpm.WorkExp, [0, 2, 5, 10, 20, 100], labels=["1-2 years", "3-5", "6-10", "11-20", "21+"])
print("   " + " | ".join(f"{band} {(g.ICorPM == 'People manager').mean():.0%}" for band, g in icpm.groupby("experience", observed=True)))

M1 how highly people rank mentors among 14 job satisfaction factors, by years of experience
   experience   mentors in top 3  average rank   mentoring others in top 3       n
   1-2 years                 19%           7.8                          5%   2,080
   3-5                       14%           8.5                          6%   3,979
   6-10                       9%           9.4                          7%   5,794
   11-20                      6%          10.1                          6%   7,360
   21+                        4%          10.6                          6%   5,192

M2 learned to code in the past year from a colleague or on-the-job training
   fully remote           26%   n=7,757
   hybrid or flexible     33%   n=11,934
   in-person              27%   n=4,584

M3 people managers among respondents, by years of experience (who could mentor you)
   1-2 years 5% | 3-5 8% | 6-10 13% | 11-20 20% | 21+ 21%

What the data says

Mentors matter most early on. Respondents ranked 15 options for what makes a tech job satisfying. With one or two years of experience, 19% put expert mentors in their top three; by 21 years, only 4% did. Mentors also climbed from an average rank of 10.6 for the most experienced to 7.8 for the newest. That fits common sense: guidance is most valuable while you are still learning how work is done.

Being in the office does not create mentoring by itself. In the past year, 27% of people working fully in person said they had learned from a colleague or on-the-job training, about the same as fully remote workers (26%). Hybrid workers were highest, at 33%. Sitting near experienced people is not enough. Someone has to ask questions and someone has to make time to answer, and that is true whether you are remote or not.

Potential mentors are a minority. Only 5% of respondents with one or two years of experience managed people, rising to about a fifth after 11 years. Most of the people you can learn from are senior individual contributors, not managers, and their time is limited. That shapes how you should ask.

What a mentor can and cannot do

Research supports mentoring, with a caveat. A 2004 meta-analysis found mentored people did better on career outcomes such as pay and career satisfaction, but the effect on objective outcomes like pay was small (Allen et al., 2004). A broader 2008 review across youth, academic and workplace mentoring found favourable outcomes of many kinds, again with generally small effects, and larger ones for workplace and academic mentoring (Eby et al., 2008).

So a mentor is not a shortcut to a promotion. What mentors do well is help you avoid wasted months: pointing you to the skill that matters, reviewing your work honestly, explaining how decisions are really made in a company, and occasionally recommending you. The work is still yours.

How to find one

1. Start with a specific question, not "will you be my mentor?"

"Could you look at my SQL for this churn analysis and tell me what you would do differently?" is easy to say yes to. "Will you mentor me?" asks for an unknown amount of time. Many lasting mentoring relationships begin with one good question and grow because the follow-up was worth the mentor's time.

2. Look close to home first

Your best candidates are people who already see your work: a senior colleague, a tech lead on the next team, a former manager, a lecturer, or the person who reviews your pull requests. Inside a company, ask your manager who knows the area you want to grow into; managers often enjoy making that introduction.

3. Then widen the circle

Outside work, the people most likely to help are those you have given something first: answers in a community forum, a useful issue or fix on an open-source project, a write-up of a meetup talk with the speaker tagged. Communities around tools you use, such as local data meetups or project discussion boards, are full of practitioners who like helping people who show effort.

4. Make it easy to keep helping you

• Do the homework first. Bring what you tried and where you got stuck.
• Keep meetings short and prepared: three questions, in writing, ahead of time.
• Act on advice and report back. "I tried your suggestion and the query went from 40 seconds to 2" is the strongest reason for someone to keep investing in you.
• Respect the end. Some mentors help for a season. Thank them and stay in touch.

5. Have several, not one

Different people help with different things: one with technical depth, one with how the company works, one with career moves. Peers a year or two ahead of you are underrated; they remember exactly what you are going through.

When you become the senior person

Interest in mentoring others was low at every experience level: only 5% to 7% ranked it in their top three. That is an opportunity. Mentoring juniors is one of the clearest signs of readiness for senior roles, as how to get promoted in a data team explains.

Related reading: networking for introverts for building the wider network, and imposter syndrome in tech for the doubts that stop many people from asking.

Structured guidance from working practitioners is part of the Data Science course.

Sources

• Career benefits associated with mentoring for protégés: a meta-analysis (Allen, Eby, Poteet, Lentz and Lima, Journal of Applied Psychology, 2004) — PubMed (US National Library of Medicine)
• Does Mentoring Matter? A Multidisciplinary Meta-Analysis Comparing Mentored and Non-Mentored Individuals (Eby et al., Journal of Vocational Behavior, 2008) — PubMed (US National Library of Medicine)
• 2025 Stack Overflow Developer Survey — Stack Overflow