Ava Tech Search

Insights.

Short, sourced notes on the US AI talent market: which roles are moving, what they pay and what is changing.

Forward deployed is the fastest-growing title in AI engineering

LinkedIn found forward deployed engineering roles grew 42-fold between 2023 and 2025, against 13-fold for AI engineer roles. Median advertised base sits around $174k, and most demand comes from seed to Series A companies that need someone who can code and stand in front of a customer.

Sources: LinkedIn labour-market data via Techscoop; Bloomberry via Fullstack; Paraform.

Inference is where the scarcity is

Serving models fast and cheaply has become its own discipline. Published ranges for inference, GPU and kernel specialists reach $300k–$500k+ in the top bands, and job ads now name CUDA, Triton, CUTLASS and FP8 directly.

Sources: Second Talent; NVIDIA and frontier-lab job postings.

AI is now the hardest skill to hire for

ManpowerGroup surveyed 39,000 employers for its 2026 Talent Shortage Survey and found AI skills are, for the first time, the hardest to find globally. 91% of organisations say hiring AI talent is a priority this year, with most focused on senior hires.

Sources: ManpowerGroup via Staffing Industry Analysts; Infragistics Reveal 2026.

Why US work authorisation matters more than it used to

Since September 2025, new H-1B petitions carry a $100,000 fee. Existing visa holders already in the US are not affected, but early-stage companies now strongly favour candidates who can start without a new petition.

Source: TechTarget.

Where the AI jobs are

The San Francisco Bay Area has the largest AI workforce in North America at about 98,700, with New York second at about 68,000. AI roles now make up 31% of US tech job listings, and 57% in the Bay Area.

Source: CBRE Scoring Tech Talent 2026.