The US AI job market, in three roles
The US AI job market is enormous and noisy. To make sense of it, focus on the three roles that build AI directly — the Machine Learning Engineer, the AI Engineer, and the Forward Deployed Engineer. They are the family tree of AI engineering, and people use the titles interchangeably even though they are not the same job, the same pay, or the same career.
Andrew Ng — the most-cited name in AI education — frames it cleanly: “As the AI Engineer role matures, I expect it to fragment into more specialized roles… AI FDEs, LLMOps Engineers, Evals Engineers.” Here is where that fragmentation stands today.
Start with the market
One window, every role measured the same way — how much its US monthly openings grew from Q1-25 to Q1-26, with the role’s volume today beside it. The shape tells the whole story: the AI Engineer is the biggest pool and still growing well; the ML Engineer is steady; the Forward Deployed Engineer is the smallest in raw count but grew the fastest by a wide margin — the signature of a role early in its curve.
High growth on a low base is not hype — it is serious commitment showing up early. The names moving first are exactly the ones you would want them to be:
“Take whatever number of people you thought might be in jobs related to AI deployment in the enterprise and multiply it by 10. Then probably 10 again.” — Aaron Levie, CEO, Box
Source: LinkedIn Jobs Opening data shared with Interview Kickstart (Q1-26 vs Q1-25 run-rate) · frontier-lab commitments per public announcements, 2026.
Same DNA, different shape
Put the three on one skill radar and they share an engineering core — Python, LLM fundamentals, RAG, production-grade code, evals. The ML Engineer goes deepest on math, modelling, and MLOps; the AI Engineer leads on LLM-app systems design and agentic orchestration. The Forward Deployed Engineer keeps the AI build and separates on exactly two axes the others barely touch — Customer Engagement and Exec Communication. That added layer is the whole role, and it is what puts the FDE at the top of the pay band.
Directional, PM-synthesized — relative shape, not a precise score. Toggle a role to isolate it.
| Machine Learning Engineer The model builder | AI Engineer The application builder | Forward Deployed Engineer The customer-embedded builder | |
|---|---|---|---|
| Focus | Training, tuning, and serving models. Data- and research-adjacent. | Building applications on top of foundation models — agents, RAG, LLM features. | AI Engineering plus customer-facing delivery — building solutions inside the customer's environment. |
| What you build | Models and ML pipelines — training jobs, feature stores, model serving, MLOps. | LLM-powered products — agentic systems, RAG pipelines, evals, guardrails, the AI features inside a product. | Everything an AI Engineer builds, but deployed and integrated in a real customer's systems — discovery to handover. |
| Core skills | ML/DL, data engineering, Python, experimentation, MLOps, cloud ML infra. | LLM APIs, prompt engineering, RAG, multi-agent (LangGraph), evals/observability, app deployment. | AI Engineering, solution architecture, customer scoping, enterprise integration, production deployment, stakeholder craft. |
| Customer-facing? | Rarely — mostly internal, works with data and product teams. | Sometimes — product-facing, occasionally customer-facing. | Always — 99% of FDE postings are customer-facing, 91% embedded. |
Wondering where your current role sits on this radar? Score yourself against the FDE bar →
What they pay, by level
Pay tracks seniority more than title. Here is the US total-comp picture across three levels — the FDE row runs widest at the top because it is bimodal: enterprise FDEs sit near the ML/SWE line, frontier-lab FDEs run far above it.
| Role | Mid · ~3–6 YOE | Senior · ~6–10 YOE | Staff+ · ~10+ YOE |
|---|---|---|---|
| Software Engineer Baseline reference. | $180K–$250K | $250K–$350K | $350K–$500K |
| Machine Learning Engineer Carries a premium over generalist SWE. | $190K–$280K | $280K–$400K | $400K–$600K |
| AI Engineer Newer title — wide spread; often folded under SWE-AI/ML. | $160K–$240K | $230K–$350K | $350K–$580K |
| Forward Deployed Engineer Bimodal — enterprise (Palantir-tier ~$211K) vs frontier-lab senior ($560K+). | $150K–$220K | $210K–$340K | $340K–$785K |
Total compensation (base + equity + bonus), US, levels.fyi 2026 — directional bands, not offers. FDE is bimodal: Palantir-tier ~$211K vs frontier-lab senior $560K+. Our own 408-JD scan reports posted BASE bands of $113K–$307K.
Want the bands employers actually post, not market averages? We decoded 408 live FDE job posts →
And AI is no longer just these three
Zoom back out. AI has crossed from a specialist skill to a baseline expectation across core tech roles. In a June-2026 scan of 1,299 US postings, nearly half of Software Engineer and Product Manager jobs — roles that are not “AI jobs” — now explicitly ask for AI, naming tools like Claude Code, Cursor, and Copilot. The three roles above are simply where AI is the core build, not a bolt-on.
Share of postings that explicitly touch AI · 1,299 US LinkedIn JDs, June 2026 (growth_engine “AI Is No Longer Optional” scan). AI-skill postings carry a ~28% salary premium (~$18K/yr, Lightcast).
IK's AI Engineering and FDE curriculum is built and refreshed against this exact market — the roles, skills, and pay bands you're reading here.
We have mapped the market and the three roles. Next, go one level deeper into the role pulling away the fastest — how deployment became its own function across every frontier lab.
AI Deployment Role Comparison →An Interview Kickstart advisor walks you through where you stand today, the exact gap to close, and the fastest route to a Forward Deployed Engineer offer — built around your background.
Book a call with an advisor →