Functional discipline.
AI Engineering Roles.
Where AI meets the codebase.
Source.
AI engineering is the largest category in this report and the one most candidates are competing for. AI engineers, ML engineers, LLM engineers, NLP engineers, computer vision engineers, application developers, integration specialists. The work is building software that calls models, evaluates outputs, handles failures, and ships into a real product. Pay scales with how recently the engineer touched a frontier model in production. Engineers with 12+ months of shipped LLM work command the steepest premiums. Roles below are sorted by base salary.
Roles in ai engineering roles.
Top cities.
Median ai engineering roles salary across the top U.S. metros for AI compensation. Multiplier reflects local cost of living relative to the national median.
San Francisco, CA
$180K
+35% vs. national
New York, NY
$170K
+28% vs. national
Seattle, WA
$165K
+24% vs. national
Boston, MA
$162K
+22% vs. national
Austin, TX
$137K
+3% vs. national
Denver, CO
$144K
+8% vs. national
Los Angeles, CA
$157K
+18% vs. national
Chicago, IL
$142K
+7% vs. national
Atlanta, GA
$136K
+2% vs. national
Washington DC, DC
$166K
+25% vs. national
Common questions.
Production LLM engineers: people who have actually shipped LLM-powered features to real users and watched what breaks at scale. Followed by retrieval engineers (RAG, vector databases, hybrid search) and tool-use engineers (function calling, agent loops). Pure ML training engineers are still in demand but the volume is smaller.
Less than people assume. Most AI engineering work in 2026 is calling APIs, building eval pipelines, designing retrieval, and shipping product surfaces. Training is its own specialty. Engineers who can do both command premium compensation, but pure API integration AI engineers are getting hired at strong rates too.
Very. Compensation gaps between top frontier labs and Series A startups can be 2 to 3x for the same nominal role. Equity packages at top AI startups often beat base plus equity at frontier labs over a 4-year vest. Total comp differences are larger than base salary differences.
Ship one real LLM-powered feature into production. Does not need to be at your job. A side project with real users counts. The signal hiring managers want is: this person has shipped, watched it fail in interesting ways, and learned. Theory without shipped work loses to shipped work without theory.
Related AI hubs.
AI Healthcare Roles
Where AI meets medicine.
AI Finance Roles
Where AI meets capital.
AI Legal & Policy Roles
Where AI meets law and policy.
AI Security Roles
Where AI meets adversaries.
AI Industrial Roles
Where AI meets the physical world.
AI Consumer & Commerce Roles
Where AI meets the consumer.
AI Creative Roles
Where AI meets craft.
AI Executive Roles
Where AI meets the C-suite.
AI Research Roles
Where AI meets the frontier.
AI Product Roles
Where AI meets the roadmap.
AI Design Roles
Where AI meets the interface.
AI Operations Roles
Where AI meets production.
Negotiate your next offer with confidence.
Open the salary calculator