Forward deployed engineers, a Palantir era specialist role, embed inside client teams for roughly 45 days at a time to ship AI features, and the new AWS unit lands as Amazon cuts more than 30,000 corporate jobs.
Amazon is cutting more than 30,000 corporate jobs. It just committed $1 billion to hire specialists who leave the office, sit inside a customer company for roughly 45 days at a stretch, and write production-grade code that turns AI models into real workflows. That contradiction is the point of the new unit Amazon Web Services announced on Tuesday.
The forward-deployed engineering team, led by AWS vice-president of frontier AI engineering and services Francessca Vasquez, will arrive in customer-company conference rooms as AWS employees and ship features from inside the client's own stack. They are not consultants in the McKinsey sense, and they are not staff augmentation. They are software engineers whose job is to make the model work in the customer's environment before they leave.
The role is older than AWS's adoption of it. Palantir Technologies built the forward-deployed model more than a decade ago, planting engineers inside client organizations to turn data and algorithms into operational software the customer actually uses. Salesforce, Anthropic, and Google Cloud have since built their own versions. AWS is now joining that pack, and it is doing so with a number attached: an initial $1 billion commitment, 5-6 small embedded teams (which AWS calls pods), and a 45-day engagement window, according to Amazon's announcement.
The unit's purpose is narrow enough to be specific. Vasquez told the Business Times in an interview prior to the announcement that customers are asking for help "to really drive agentic AI patterns in their workflows," meaning software that lets an AI model act on its own rather than just answer. A pod is meant to install that capability inside the customer's own systems in under two months, faster than the multi-quarter projects enterprise software sales used to require.
The numbers come from different parts of Amazon. The parent company has cut more than 30,000 corporate roles since October, CNBC reported, as part of a broader pullback across the technology sector. AWS is now spending $1 billion on a hiring push for a specialist that almost nobody outside the industry has heard of. The two decisions are not unrelated: enterprise software is being sold less like a license and more like a deployment, and the engineer who embeds with the buyer has become the unit of account for that shift.
The labor market is already pricing the role in. Job postings for forward-deployed engineers and adjacent AI-deployment roles have grown more than 700% over the past year, according to Business Insider, citing aggregator data. OpenAI and Google are racing to hire the same kind of engineer, per The New Stack, and the skills the role rewards are specific: production-grade coding, the patience to navigate a customer's internal politics, and the taste to know which agentic AI workflow is worth building first.
What AWS is buying with its $1 billion is not really a hiring announcement. It is a position in a market where the hyperscaler that can ship inside the customer's stack wins the deal, and where the labor market for that skill is being repriced almost weekly. The watch item for the rest of the year is whether 45-day pod cycles are short enough for AWS to keep up with Palantir's decade of muscle memory, or whether the unit ends up as a press-release footnote next to another quarter of broader enterprise-AI competition.