Cognizant's Project Leap: The 12,000 Aren't the Product Engineers
Cognizant's Project Leap will cut 12,000 to 15,000 jobs, mostly in India. Here's why the displaced pool isn't product talent, and where it actually fits.
Coverage of Cognizant's Project Leap has flattened a 357,000-person services firm into a single headline: "12,000 to 15,000 engineers laid off." If you run sourcing for a product org, that framing is going to waste your week. The displaced pool is not the pool you think it is, and treating it like a generic tech layoff wave will fill your pipeline with the exact profiles AI is automating, not the ones product teams want to hire.
Here is what the Cognizant Project Leap layoffs actually represent, and where in a sourcing pipeline they belong.
What Project Leap actually is
On April 29, Cognizant disclosed it expects $230M to $320M in severance-related costs under Project Leap: $200M to $270M in employee severance and personnel expenses, plus $30M to $50M in other restructuring. The reported scope is 12,000 to 15,000 jobs globally, with the heavy concentration in India. Moneycontrol's math (Rs 15 lakh average annual salary, ~6 months severance, Rs 7.5 lakh per head) lands the India impact alone at roughly 12,000 to 13,000 people.
This is not a distressed company shedding cost. In Q1 2026, Cognizant posted $5.4B in revenue, up 5.8% YoY, with trailing 12-month bookings of $29.6B and a headcount of 357,600 as of March. They are also hiring 20,000+ freshers in 2026. Cuts and hires at the same time is the tell. This is a pyramid reshape, not a contraction.
CEO Ravi Kumar S has been explicit about the model: a "broader and shorter pyramid" combining digital tools, automation, and human workers as AI absorbs more service delivery. CFO Jatin Dalal framed Project Leap to analysts as a "cost of delivery" exercise. Translation: the middle of the pyramid is being thinned, freshers are being hired cheaper, senior architects are being kept.
The roles being cut tell you who's in the pool
The exposed work categories are specific, and they matter for how you triage inbound resumes:
- Entry-level software testing and manual QA
- Code maintenance on legacy stacks
- Tier-1 application support
- Traditional database administration
- Repetitive back-office processing
- "Standard development work" on long-running client engagements
The experience band most at risk, per reporting on the cuts, is 6 to 12 years in routine roles. That is exactly the band recruiters often mistake for "senior" when they scan a resume. Twelve years at Cognizant doing tier-1 support for an insurance client is not twelve years building product. It is twelve years inside a billable utilization model, where the incentive was hours booked against a statement of work, not features shipped to users.
This is also why AI is doing the cutting. The work being automated was already low-judgment by design. The pyramid model required it to be: armies of freshers handling repeatable tickets so the firm could bill the client. The same characteristics that made the work staffable in 2014 make it automatable in 2026.
Why "12,000 engineers" misleads sourcers
Plug "ex-Cognizant" into LinkedIn and you get a wall of people whose titles read Software Engineer, Senior Software Engineer, Technical Lead. The titles do not distinguish between someone who shipped a payments rewrite and someone who patched COBOL batch jobs on a maintenance contract. Boolean search cannot see the difference. Title-based outreach to the Project Leap cohort will land you a high response rate and a low conversion rate, because the response rate is driven by urgency, not fit.
This is the friction we built Refolk for. You describe the engineer you actually want ("backend engineer who has shipped greenfield Go services at a product company in the last three years, not maintenance work at a services firm") and get a ranked shortlist instead of 4,000 same-title profiles to manually filter.
The CEO's quote is the actual sourcing signal
"Broader and shorter pyramid" is not corporate filler. It is a literal description of what Cognizant is keeping versus cutting.
- Keeping and growing: senior architects, AI-fluent engineers, client-facing tech leads at the top of the pyramid. Freshers at the bottom.
- Cutting: the 6-to-12-year middle that built its career executing inside the old model.
If you are sourcing India engineering talent for a product role, the high-signal Cognizant alumni are mostly the ones who already left in the last two or three years, before Project Leap. They self-selected out of the pyramid into GCCs, product startups, or US-headquartered captives. The Project Leap cohort, by definition, did not.
The strongest Cognizant alumni for product roles are the ones who left two years ago, not the ones being cut next quarter.