The embedded-engineer model has become a hyperscaler race. AWS put a billion dollars behind it, Microsoft branded a whole business around it, and Google is now building the same thing, just far more quietly. There is no Google press release and no dollar figure. The evidence is in Google’s own job listings, a named team that appeared on stage at Google Cloud Next, and its cloud chief saying so out loud.

A team working at dark teal-lit consoles in a control room, representing engineers operating a system together.
Forward deployed engineering puts a provider's own builders inside the customer's environment. Google is now hiring for exactly that role.

What happened

Google is assembling a forward-deployed-engineering function inside Google Cloud. The primary evidence:

  • Its own job listings. Google’s careers site advertises a “Forward Deployed Engineer” job family across seniority levels (GenAI and Applied AI, Google Cloud), and an explicitly titled “Forward Deployed Engineering” team. The role is described as an embedded builder who codes, debugs, and ships bespoke agentic solutions inside the customer’s environment, the same pattern as AWS and Palantir.
  • A named team, already operating. At Google Cloud Next 2026 (22 April 2026), Google presented a “manager of applied AI forward deployed engineering,” and CEO Thomas Kurian’s keynote framing was “the era of the pilot is over, the era of the agent is here.”
  • The CEO’s stated intent. Kurian said publicly that Google is “investing in hiring additional forward-deployed engineers to help us scale customer AI transformation.”

What Google has not done is announce this as a packaged program. There is no press release, no disclosed investment, and no official headcount. Reporting (The Information) puts the hiring in the hundreds, but Google has not confirmed a number, so treat that as unverified. This is also genuinely new, not just Google’s long-standing “Customer Engineers” or Google Cloud Consulting renamed: the explicitly-titled Forward Deployed Engineer job family and team are the new element.

How it compares

AWSMicrosoftGoogle
WhatForward Deployed Engineering unitFrontier CompanyForward Deployed Engineering team
How announcedPress release (June 2026)Company blog (July 2026)Job listings, CEO statements, Cloud Next
InvestmentUSD 1 billionUSD 2.5 billionNot disclosed
HeadcountGrowing to thousands6,000 embeddedReportedly hundreds (unconfirmed)
Model5-6 engineer pods, 45-day engagementsEmbedded deliveryEmbedded builders in customer environment

Why it matters for builders

Three of the largest cloud providers have now reached the same conclusion: the bottleneck to enterprise AI is not the models, it is the shortage of people who can turn them into working, governed systems. When AWS, Microsoft, and Google all start shipping their own engineers into customer teams, “who builds it” becomes a product.

For independent engineers, agencies, and consultancies, that is both a threat and a map. The hyperscalers will compete on scale and platform depth. The room they leave is what they do not do well: independence from a single vendor’s stack, deep domain knowledge, honest architecture advice, and cross-cloud flexibility. Either way, the demand signal is unmistakable, and it favours anyone building the skill of taking AI from pilot to production. See the AWS and Microsoft moves for the rest of the pattern.

Sources

Further reading