Google Rehired 1 in 5 of Its AI Engineers. Alumni Is a Channel Now.

Boomerang hiring is no longer an HR curiosity you tolerate. In March 2025, ADP payroll data recorded the highest boomerang share since it started tracking in 2018, and the information sector hit roughly two-thirds. If "alumni" isn't a named channel on your sourcing plan next quarter, you're sourcing below the market floor.

Google's 20% is the headline. It's also below the market.

Twenty percent of Google's 2025 AI software engineer hires were boomerangs, per CNBC's December 2025 reporting, and Sergey Brin has personally reached out to some of them. That sounds like a lot until you compare it to the rest of the economy.

ADP's payroll data puts March 2025 boomerang share at 35% of all US new hires, up from 31% in March 2024 and 26% at the March 2022 Great Resignation trough. In the information sector specifically, roughly two-thirds of March 2025 new hires were returning employees, roughly double the year before. Nela Richardson, ADP's chief economist, called it "the highest March for boomerang hires ever," reflecting "a labor market that's very stable, but also with hiring that's very cautious."

The mechanism is straightforward. The 2022 to 2023 layoff cycle seeded a two-year reservoir of alumni, remote work removed relocation as a blocker, and hiring managers who slashed heads are now cautiously refilling from the pool they know. Google didn't invent this pattern. It's a mainstream, measurable sourcing channel that most TA orgs still treat as a referral fluke.

66%

The 2023 layoff cohort is the pool feeding today's rehires.

Every mass rehire cycle in 2025 traces back to a specific layoff cohort in 2022 or 2023, and the biggest one is Google's. Approximately 12,000 Google employees were cut globally in early 2023, including specialists in machine learning, cloud computing, and large language models. That's the alumni pool Brin is dialing.

The numbers behind Google's cost stack tell the real story:

• 2023 severance: 16 weeks base plus 2 weeks per year of tenure.
• 2023 to 2025: roughly a two-year productivity gap while that talent worked elsewhere.
• 2025 rehire cost: reports of a 25% salary premium, often with title bumps and promotions.

That is the "boomerang premium," and it's the number to put in front of any hiring manager who defaults to "no rehires." Refusing to rehire doesn't save money. It converts a temporary cost cut into a permanent wage tax, because the skill still gets bought on the open market at 1.25x the original comp.

Refusing to rehire doesn't save money. It converts a temporary cost cut into a permanent wage tax.

In Refolk's index, ex-Google AI engineers are already boomeranging.

The cycle is visible directly in the data. In Refolk's index of professional profiles, a base search on US-based ML and AI engineers with Google work history returns 363 matching profiles, and the top current employers of that cohort are Google (9), Meta (6), Google DeepMind (3), Pinterest (2), and Uber (1).

Read that carefully. Of the top five current employers of ex-Google AI engineers, two are Google itself. Twelve of the top-25 sampled employers, roughly 48% of the top-employer distribution, are Google or its DeepMind subsidiary. The boomerang cycle isn't a projection. It's already happening at measurable volume, and it often means returning to DeepMind rather than literally the same team.

Geography reinforces the point. Top regions for this cohort are the San Francisco Bay Area, New York City, Mountain View, and Sunnyvale. These are the same zip codes the roles sit in, which means alumni sourcing sidesteps the relocation and visa friction that kills external funnels. That is the exact gap Refolk closes for alumni work: describe "ML engineers with Google or DeepMind history, currently outside Google, in the Bay Area or NYC" in plain English and get a ranked shortlist without a Boolean session.

Alumni beats external in three ways external can't fix.

Alumni sourcing wins because it starts with candidates who are already technically qualified, culturally screened, and geographically placed, which no external funnel can replicate. Three concrete advantages:

1. Ramp is compressed. They know the stack, the codebase conventions, the team norms, and often the same manager. First-90-day productivity looks nothing like an external hire's.
2. Screening cost is near zero. You have their performance history, calibration data, and 360 feedback. Interview loops can be shortened or skipped for senior ICs.
3. Geography is pre-qualified. LinkedIn and ADP both flag remote work's geographic decoupling as a key enabler. Alumni who relocated in 2021 or 2022 can now rejoin without another move.

The employer-side sentiment has caught up. LinkedIn's 2024 Workplace Report found boomerang hires grew 35% since 2022, and 68% of HR leaders reported they were more open to rehiring former employees than ever before. The Workforce Institute at UKG surveyed 1,800-plus HR professionals and found 76% now say they are more accepting of rehiring former employees. The stigma is dead at the leader level. It survives mostly as an untested reflex among individual hiring managers, which is a coaching problem, not a policy one.

The optimal outreach window is 18 to 24 months post-departure.

Time the outreach to the disillusionment curve, not the departure date, and the reply rate roughly doubles. Six months is too early: the new job still smells fresh, the signing bonus is unvested, and the story they told friends about leaving is still load-bearing. Forty-eight months is too late: skills have drifted, comp expectations have reset, and they've built new manager loyalty.

The 18 to 24 month window matches the Google 2023 to 2025 cycle almost exactly, and it maps to a specific psychological pattern in boomerang-motivation data. About two-thirds of returning employees cite better work/life balance as a motivating factor, and reporting consistently flags that the new employer's stated flexibility often turned out to be nominal. That gap between promised and delivered flexibility takes 12 to 18 months to fully register.

Practical outreach signals to look for inside that window:

• Title increase at the new employer of one level or less (still hungry, not entrenched).
• A recent RTO announcement at the new employer.
• Team reorg or manager change on their side (loyalty anchor is gone).
• Public GitHub or blog activity that's slowed in the last quarter (engagement dropping).
• Second layoff round at the new employer that they survived (survivor guilt is a signal, not just for the cut).

Founder outreach converts. Recruiter InMail doesn't.

For senior and staff-level alumni, InMail from a recruiter has near-zero conversion, and a note from a former manager or founder converts at multiples of that. This is why Brin is dialing personally, and why Mark Zuckerberg has reportedly reached out to researchers directly for Meta. It's not vanity. It's an acknowledgment that at the senior end, the message is the funnel.

A workable playbook for TA leaders who aren't founders:

• Tier 1 (staff and above): route to the original hiring manager or a peer VP, with a two-sentence brief you write. Recruiter never touches the first message.
• Tier 2 (senior IC): route to a former teammate still at the company for a "coffee, no pitch" note. Recruiter joins on message three.
• Tier 3 (mid-level): recruiter outreach is fine, but reference the specific team and problem, not the generic req.

The bottleneck is candidate identification, not message-drafting. Most alumni CRMs are stale within 90 days because they don't track where the person went next. That is exactly where a plain-English search across GitHub, LinkedIn, and the open web pays off: ask Refolk for "everyone who left our Search Infra team between 2022 and 2023, currently at Series B startups," and skip the manual LinkedIn reconciliation. Vendors like Enterprise Alumni handle the community and content side well, but they're not built for surfacing the "where are they now" signal at the resolution a sourcer needs.

Build the alumni channel like a proper pipeline, not a database.

Treat alumni as a channel with a target, a funnel, and a monthly review, the same way you'd treat referrals or inbound. The teams doing this well have four things in place:

1. A named owner. One recruiter or sourcer owns alumni, part-time or full-time depending on company size. If it's everyone's job, it's no one's.
2. A refreshed list, quarterly. Ex-employees move. You need current employer, current title, and current location refreshed at least every 90 days. This is the step most alumni programs skip and where Refolk becomes load-bearing: you re-run the same plain-English query each quarter and get an updated shortlist without rebuilding Boolean strings.
3. A "do not contact" list, honestly maintained. Some alumni left for real reasons. Note them, respect it, and move on. This protects the brand more than any nurture campaign.
4. A comp policy for rehires. Decide in advance whether you're paying the boomerang premium, matching current market, or offering equity in place of cash. Ambiguity here kills more offers than the number itself.

The information-sector companies hitting 45% boomerang share over the trailing 12 months aren't doing anything exotic. They're just treating alumni as a first-class channel while the median TA org still treats it as a warm-lead spreadsheet.

FAQ

Is a 20% boomerang rate healthy or a warning sign?

For an information-sector company post-layoff, 20% is fine but unremarkable. The March 2025 all-industry average was 35%, and info-sector companies averaged 45% over the trailing 12 months. If you're in tech and running under 25% from alumni, you're leaving warm, pre-screened candidates on the table and paying the external-hire premium instead. If you're over 60%, check whether you're overfishing a shrinking pool.

How do I overcome a hiring manager's "no rehires" reflex?

Reframe it as a comp question, not a loyalty question. Google is reportedly paying a 25% salary premium to bring back employees it laid off two years earlier, often with title bumps and promotions. That premium exists whether you rehire the person or not, because the same skill trades at that price on the open market. The choice isn't "boomerang vs no boomerang." It's "boomerang at market vs external at market plus ramp cost."

What's the right time window to reach out after someone leaves?

Eighteen to twenty-four months post-departure hits the sweet spot for most senior ICs. Six months is too early (still honeymooning), forty-eight months is too late (roots grown). Watch for signal events inside the window: RTO announcements, manager changes, or a second layoff at the new employer. Two-thirds of returning employees cite work/life balance as a motivator, and that disillusionment typically registers in the 12 to 18 month range.

Should recruiters or executives own alumni outreach?

Split it by seniority. Staff and above should get outreach from a former manager, peer VP, or founder, with the recruiter staying invisible until message three or four. Senior ICs get a former teammate as first-touch. Mid-level roles are fine for direct recruiter outreach as long as the message references the specific team and problem rather than a generic req. Brin and Zuckerberg reaching out personally isn't celebrity theater. It's the only outreach mode that converts at the senior end.