What each role actually does, and which tools show up in the interview.
AI Engineer
Ships product features on top of foundation models: retrieval, evaluation, tool calling and latency/cost tuning.
- Python or TypeScript
- RAG & vector search
- Prompt + eval design
- API/infra work
Typically posted at $130k–$250k in the US, £70k–£130k in the UK
Tools to know: Cursor, GitHub Copilot, Claude
Prompt / Applied AI Specialist
Owns prompt libraries, evaluation datasets and guardrails for a product team. Often sits between product and engineering.
- Evaluation methodology
- Domain expertise
- Data analysis
- Model comparison
Typically posted at $90k–$160k
Tools to know: ChatGPT, Claude, Perplexity
ML Research Engineer
Trains, fine-tunes and evaluates models. Requires deeper math and distributed-training experience than an AI engineer role.
- PyTorch
- Distributed training
- Paper reproduction
- CUDA basics
Typically posted at $180k–$400k+ at frontier labs
Tools to know: GitHub Copilot, Claude
AI Content & Growth
Uses AI tooling to scale content, SEO and lifecycle marketing while keeping human editorial review in the loop.
- SEO
- Editing
- Workflow automation
- Analytics
Typically posted at $60k–$120k
Tools to know: Jasper, Grammarly, Perplexity