The Skills Are Not Evenly Short
“Talent shortage” is too blunt to act on. Scarcity in India is concentrated in four places, and knowing which one your role sits in changes what you should do about it.
Where the gap is, on the published data
These figures describe the Indian market. The live panel further down is our own desk.
Figures published by third parties and cited above. They describe the Indian market, not WSNE’s own performance.
The five genuinely short pools
Everything else in engineering is competitive rather than scarce — which matters, because the two need different strategies and only one of them justifies the premium.
AI, GenAI and machine learning
The deepest gap in the market. Demand up over 300% since 2024 against a ~53% skills deficit — and the hardest pool to screen, because the gap between people who have shipped models to production and people who have completed courses is very wide.
How we recruit for thisCloud, platform and reliability engineering
Kubernetes and Terraform at genuine production depth remain persistently short. Widely claimed on CVs, far less widely held. This is the largest skill cluster in our own mandate records.
How we recruit for thisData engineering
BFSI GCCs alone report a 42% data skills gap. Pipeline and warehouse engineers who can own reliability, not just build transformations, are the specific shortage.
How we recruit for thisSecurity architecture and offensive security
SOC analysts are available; people who can design a security posture or run genuine offensive testing are not. Seniority, not headcount, is the constraint.
How we recruit for thisTest automation with engineering depth
Manual QA is plentiful. Engineers who can build and own an automation framework rather than write scripts inside someone else’s are consistently hard to find.
How we recruit for thisWho you are competing with
At the senior end of the pool, increasingly this.
People added by Indian GCCs — close to twice what IT services firms hired
Source: The Intech GroupFigures published by third parties and cited above. They describe the Indian market, not WSNE’s own performance.
How we run a scarce-skill mandate
Scarce roles are not ordinary roles worked harder. The pool is small enough that approach quality matters more than volume, and the wrong first message removes a candidate permanently.
Before you approve the requisition
Four questions that decide whether this role takes six weeks or six months:
- Does the work need this specific tool, or the capability underneath it?
- Is the band benchmarked to this year, in this city, for this scarcity?
- Could a strong adjacent candidate be productive within a quarter?
- Is the start date planned from a 65–80 day notice period, or from hope?
Questions employers ask
Which tech skills are genuinely scarce in India in 2026?
Scarcity is concentrated rather than general. Demand for AI specialists has grown over 300 per cent since 2024 against an estimated 53 per cent skills deficit, and BFSI global capability centres report a 42 per cent gap in data skills specifically. Beyond AI and data, the persistent shortages sit in cloud and platform engineering — Kubernetes and Terraform at genuine production depth — and in security architecture and offensive security. General application development is not scarce; senior application development with production ownership is.
Why do niche roles cost so much more?
Because a small pool receiving many offers sets its own price. Published figures put the premium for genuinely niche candidates at around 1.7 times what adjacent roles command. The mistake employers make is treating that premium as negotiable — it is a market rate, not an ask. The negotiable question is whether the role truly needs the niche skill, and often it does not.
How long does it take to hire an AI or data engineer in India?
Longer than the search suggests, and the delay is at the ends rather than the middle. Sourcing a genuinely qualified shortlist takes two to four weeks in a metro. Then the candidate is running several processes, so the offer stage is competitive, and senior notice periods of 65 to 80 days follow. Planning on three months from approval to a person actually starting is realistic; planning on six weeks is how roles get reopened.
Should we widen the requirement or wait for the exact profile?
Widen it far more often than employers do. A meaningful share of the briefs we receive specify a named tool where the underlying capability is what matters — an engineer strong in one cloud platform will be productive on another inside a quarter, and a Python data engineer is not disqualified by having used Airflow rather than the tool you happen to run. Relaxing one adjacent requirement frequently doubles the pool at no cost to delivery. We will tell you which requirement to relax, and which one is genuinely load-bearing.
Are GCCs making this worse for other employers?
For the senior end of the pool, measurably. Indian GCCs added roughly 200,000 people, close to twice what IT services firms hired, and they hire for product, R&D, security architecture and AI — the same profiles most employers are competing for, generally at above-services compensation. If you are hiring senior engineers in Bangalore, Hyderabad or Pune, you are competing with GCC budgets whether or not a GCC appears on your shortlist.
Have a role nobody can fill?
Send it with the band and the must-haves. We will tell you which requirement is costing you the pool before we start.