Skill pair
LLMs and RAG: demand as a pair
How often remote tech listings ask for LLMs and RAG together, which roles want the combination and what usually comes with it.
What the pairing tells you
22 positions ask for both LLMs and RAG. That is 26.5% of the 83 LLMs positions and 57.9% of the 38 RAG positions. The relationship is asymmetric. About half of RAG listings (58%) also ask for LLMs, but only about one in four LLMs listings (27%) ask for RAG. For someone who already knows RAG, LLMs is a natural next step; the reverse is a weaker bet.
Measured against the 1,265 skill-tagged positions, the pair appears 8.8x as often as it would if the two skills were requested independently, a strong affinity. Most of these positions (68%) are in Data Science & ML.
The skills most often added to this pair are Python (86%), AWS (41%) and TypeScript (41%). Together they form a recognizable three-skill bundle.
Of the 18 positions with a stated level, 89% are senior or above (sample-wide: 84%).
Overlap
Positions asking for one, the other, or both
Third skills that complete the bundle
Share of the 22 positions that ask for both
Role families
Where the pair is requested
Seniority
Example titles
- AI Engineer - Samsara
- Senior AI Engineer - ClassWallet
- Software Engineer - Ascensus
- Staff AI Engineer - Drata
- AI SWE Summer Intern - Experian
- Staff AI/ML Engineer - Dragos
- Senior AI/ML Engineer - Dragos
- Full Stack AI Engineer - Greystar
- Senior ML Engineer (ML/AI) - Lyra Health
- Applied AI Product Engineer - VetsEZ
- Senior AI Engineer, Quality - Fieldguide
- People Analytics AI Engineer - Samsara
Employers in the sample
- Dragos 2
- Drata 2
- Samsara 2
- Ascensus 1
- Bain & Company 1
- ClassWallet 1
- Enigma 1
- Experian 1
- Fieldguide 1
- Greystar 1
Questions
How often do remote listings ask for both LLMs and RAG?
22 positions in the sample ask for both: 26.5% of LLMs positions and 57.9% of RAG positions.
Should I learn LLMs and RAG together?
The data supports it: the pair appears 8.8x as often as independent demand would predict.