A global capability centre — a company's own offshore entity, staffed with its own employees, doing product, engineering or operations work. It differs from outsourcing in that the people are yours, the roadmap is yours, and the centre is intended to own outcomes rather than execute tickets.

Who should a GCC hire first?

The centre lead, before anyone else. This person sets the hiring bar, and every subsequent hire is filtered through their judgment. Hiring engineers before there is someone senior enough to define the bar is the most common and most expensive sequencing mistake.

How long does it take to hire a GCC leadership team in India?

Plan on 60–90 days for a centre lead and 45–60 for senior engineering leadership. Notice periods of 60–90 days sit on top of that, so a leadership hire signed in January realistically starts contributing in April.

Why do GCC hires fail?

Most often because the centre was sold as ownership and staffed as a delivery arm. Senior Indian engineers have seen this before and will leave within a year if the work turns out to be maintenance. The second reason is hiring for a headcount plan rather than a scope.

Should a GCC use an agency or hire in-house?

The first ten hires almost always need external specialists, because there is no internal TA function yet and the leadership roles are the hardest in the market. Build the in-house team around hire five or six, and hand over once the bar is set and demonstrated.

Building a GCC in India: The First Ten Hires

The first ten hires decide what a GCC becomes. Get the sequence right and you have a centre that owns products. Get it wrong and you have an expensive delivery arm with attrition problems, and you will spend two years unwinding it.

The mistakes are consistent and they are all made in the first six months.

Hire the centre lead first, and do not compromise

Everything downstream is filtered through this person's judgment. They set the bar, they are the reason a strong engineer takes the call, and they are who a candidate is really evaluating in the final round.

The failure mode is hiring a delivery manager for a role that needs an engineering leader. A centre lead who has run delivery but never owned a product will build a delivery organisation, because that is what they know how to build. That is not recoverable by hiring good engineers under them.

What to actually test: have they built a team from zero rather than inherited one; have they owned a roadmap rather than a backlog; can they say no to headquarters. That last one matters more than it sounds. A centre lead who cannot push back becomes a ticket queue within a year.

Budget 60–90 days for this search, plus notice.

Hires two to four: the architecture

These are your senior individual contributors — staff-level engineers who will own the core systems. Hire them before the managers.

The reason is simple: a manager without strong ICs recruits more managers. A strong IC sets a technical bar other strong engineers want to be measured against, and they become your best recruiting asset. The first excellent engineer you hire brings the next three.

Look for people who have owned a system in production at scale, not people who have worked adjacent to one. The interview question that separates them is not about design — it is "tell me about something you built that broke, and what you changed afterwards."

Hires five and six: the first manager and the first recruiter

Now hire your engineering manager, into a team that already exists.

And hire your first in-house recruiter here, not later. This is the hire most GCCs defer and most regret deferring. By hire ten you are running continuous hiring, and a recruiter who has watched the first nine searches understands the bar in a way that no brief can transfer.

Hires seven to ten: the delivery core

Now you scale — mid-level engineers, QA, whatever the product needs. This is the easiest part of the sequence, and it is the only part most headcount plans actually budget time for.

If the first six are right, these hires close quickly, because you now have a real team to sell and senior people who will vouch for it.

What makes GCC hiring hard in India

The best people are not looking. Every strong senior engineer in Bengaluru or Hyderabad already has a job, and probably a good one. They are reachable, but only through someone who can have a credible conversation about scope and ownership.

Your brand may not exist locally. A well-known Western company can be unknown to Indian engineers, or known only as "an offshore centre" — which is exactly the association you are trying to avoid. The pitch has to be built, not assumed.

GCCs compete with each other, not with services firms. Microsoft, Amazon, Google, Salesforce and several hundred others are hiring the same people. Your comp needs to be benchmarked against them, not against a services company.

Notice periods stack. Sixty to ninety days is normal. A hire signed in January starts in March or April. Any plan that does not build this in is already late.

The mistakes that are expensive to undo

Selling ownership and staffing execution. Senior Indian engineers have been burned by this and will leave inside a year. If the roadmap genuinely lives at headquarters, say so and hire people who want that — do not sell autonomy you cannot deliver.

Hiring to a headcount plan rather than a scope. "Twenty engineers by Q3" produces twenty engineers and no clarity. Hire to what the centre will own.

Lowering the bar for hires seven to ten because you are behind. Every weak hire in the first ten costs you a strong one later, because strong engineers assess the team they are joining.

Not measuring offer-to-joining. In India this is the metric that tells you whether your pitch is real. If a meaningful share of accepted candidates never arrive, the problem is not recruitment — it is what they heard between acceptance and their start date.

The sequence, compressed

Centre lead — 60–90 days, no compromise 2–4. Senior ICs who own the architecture
First engineering manager, into an existing team
First in-house recruiter 7–10. Delivery core

Use external specialists for one through six, because you have no TA function yet and these are the hardest roles in the market. Build in-house around five and six, and hand over once the bar is set and has been demonstrated on real hires.