We are not going to publish a stipend table, because we cannot verify one. What we can offer is a method that does not depend on someone else's unverifiable numbers.
Why published benchmarks mislead
Stipend data is self-reported, heavily skewed by which employers respond, and rarely segmented finely enough to be actionable. A single national "average intern stipend" mixes a Bengaluru product-company intern with an unpaid one, and means nothing.
Benchmark against your own funnel
Your offer-acceptance rate is the only stipend benchmark that is genuinely about your situation. If acceptance is high and time-to-fill short, you are competitive; if candidates decline citing other offers, you are not — and no table tells you that faster.
Segment by category, not by seniority
Early-career scarcity varies enormously by domain. A single stipend band across all functions systematically overpays the abundant categories and underpays the scarce ones.
Remember stipend is a threshold, not a lever
Below a credible level nothing else compensates. Above it, marginal increases buy much less than process speed, work specificity and a believable conversion story.
Key takeaways
- Published stipend averages mix incomparable situations and are rarely actionable.
- Your own offer-acceptance rate is the most reliable competitiveness signal.
- Segment stipend by domain scarcity, not by a single company-wide band.
- Stipend is a threshold; above it, other factors dominate.
Questions
Do you publish stipend data?+
No. We publish figures we can count from our own database — candidate counts, skills, cities, batches. Stipend data would be self-reported and unverifiable, so we leave it out rather than dress it up.
Should unpaid internships be considered?+
They shrink your pool sharply and skew it towards candidates who can afford to work unpaid, which is a composition problem as well as an ethical one.