Every salary guide has the same weakness: it reports what people say they earn, or what employers say they will pay. This one reports what was actually signed. The figures below come from 4,100 offers Memsolos placed or advised on between August 2025 and July 2026, all of them accepted, all of them verified against the offer letter.
Two things stand out this year. First, the premium for genuine platform and reliability engineering has widened again, while generalist full-stack roles have flattened almost completely. Second, the Hyderabad discount against Bengaluru has narrowed to under seven percent at senior levels, which changes the location arithmetic for anyone planning a new site.
How to read these numbers
All figures are fixed annual compensation in lakhs of rupees, excluding equity, joining bonuses and variable pay. The median is the point where half of accepted offers sat above and half below. The 75th percentile is what you should expect to pay for a candidate the market genuinely competes for.
If you are budgeting a requisition, use the median for a solid hire in a healthy market and the 75th percentile for a role you have already failed to fill once.
Engineering: individual contributors
| Role | Experience | Median | 75th percentile |
|---|---|---|---|
| Software Engineer | 2–4 yrs | 14.0 | 19.5 |
| Senior Software Engineer | 5–8 yrs | 28.0 | 38.0 |
| Staff Engineer | 8–12 yrs | 46.0 | 58.0 |
| Principal Engineer | 12+ yrs | 62.0 | 81.0 |
| Site Reliability Engineer | 5–8 yrs | 33.0 | 44.0 |
| Platform / Infrastructure Engineer | 6–10 yrs | 38.0 | 51.0 |
| Mobile Engineer (iOS / Android) | 4–7 yrs | 24.0 | 32.0 |
| QA Automation Engineer | 4–7 yrs | 17.5 | 23.0 |
Data, AI and security
| Role | Experience | Median | 75th percentile |
|---|---|---|---|
| Data Engineer | 4–7 yrs | 26.0 | 35.0 |
| Senior Data Engineer (streaming) | 6–9 yrs | 36.0 | 47.0 |
| Analytics Engineer | 3–6 yrs | 20.0 | 27.0 |
| Machine Learning Engineer | 4–8 yrs | 34.0 | 49.0 |
| ML Platform Engineer | 6–10 yrs | 45.0 | 62.0 |
| Data Architect | 10+ yrs | 52.0 | 68.0 |
| Application Security Engineer | 5–8 yrs | 31.0 | 42.0 |
| Security Architect / GRC Lead | 10+ yrs | 48.0 | 63.0 |
Engineering leadership
| Role | Team size | Median | 75th percentile |
|---|---|---|---|
| Engineering Manager | 6–12 | 48.0 | 62.0 |
| Senior Engineering Manager | 15–30 | 66.0 | 84.0 |
| Director of Engineering | 40–80 | 92.0 | 124.0 |
| VP Engineering / Site Head | 100+ | 135.0 | 190.0 |
Location multipliers
Apply these to the national medians above. They are derived from matched-role comparisons, so they control for seniority mix rather than simply averaging city payrolls.
| City | Mid-level multiplier | Senior multiplier | Note |
|---|---|---|---|
| Bengaluru | 1.00 | 1.00 | Reference market |
| Hyderabad | 0.94 | 0.93 | Gap narrowing fast at senior levels |
| Pune | 0.92 | 0.90 | Strong for platform and embedded |
| NCR (Gurugram / Noida) | 0.97 | 0.99 | Product and fintech premium |
| Chennai | 0.88 | 0.86 | Deep in enterprise and QA |
| Tier-two hubs | 0.74 | 0.71 | Thin at staff level and above |
Five things the data actually tells you
- Specialisation pays, breadth does not. The gap between a senior full-stack engineer and a senior platform engineer at identical experience is now 36% at the median. Two years ago it was 19%.
- The Hyderabad discount is disappearing at the top. At staff level and above, the differential against Bengaluru is under seven percent. Location arbitrage still works for mid-level hiring; it no longer works for leadership.
- Counter-offers have become the default. Forty-one percent of our accepted candidates received a counter-offer. The ones who still left cited scope and manager quality far more often than money.
- Notice periods are the hidden cost. Median notice in Indian technology is now 68 days. A four-week search followed by a ten-week wait is still a fourteen-week vacancy, budget for it.
- The 75th percentile is not a stretch, it is the market. Roles budgeted at the median took 2.4 times longer to fill than roles budgeted at the 75th percentile. The cheaper band is usually the more expensive decision.
Want your own roles benchmarked?
We run role-level benchmarking against live offer data for clients and prospective clients alike, usually within three working days. Send the job specifications and we will send back the bands, the supply picture and an honest read on whether the role is fillable as written.
Methodology
Sample: 4,100 accepted offers across 340 client organisations, August 2025 to July 2026. Roles were normalised to a common levelling framework before aggregation, and outliers beyond three standard deviations were excluded. Equity, variable pay and joining bonuses are excluded throughout because their treatment varies too widely to compare honestly.
Figures are indicative and should inform a band, not fix one. Compensation for any individual hire depends on scope, competing processes and how badly the market wants that person this quarter.
