How is skill demand measured?

We scan the title and description of every active AI job posting for a curated vocabulary of skills, then rank each skill by the number of distinct live roles that mention it. Share is that count divided by all active AI roles.

Which skills are counted?

Programming languages, ML frameworks, AI techniques (RAG, fine-tuning, agents), cloud and data tools, and current LLM tooling (LangGraph, LlamaIndex, vLLM, and more). Skills appearing in fewer than 10 live roles are omitted to avoid noise.

It reflects live listings and updates as new roles are posted and old ones close. Each skill links to its current openings.

Skills

In-Demand AI Skills

Python is the most in-demand AI skill, asked for in 27% of 21,148 active AI roles. The top five are Python, Machine Learning, Agentic AI, LLMs, and GCP.

Ranked by live-role demand across active AI listings. Updated 13 hours ago.

Top 15 skills by active roles

Full ranking

#	Skill	Roles	Share
1	Python	5,777	27%
2	Machine Learning	5,042	24%
3	Agentic AI	3,967	19%
4	LLMs	3,347	16%
5	GCP	2,456	12%
6	GPU	2,404	11%
7	AWS	2,386	11%
8	C++	2,181	10%
9	Kubernetes	2,114	10.0%
10	Speech Recognition	2,010	9.5%
11	CI/CD	2,004	9.5%
12	Generative AI	1,631	7.7%
13	TypeScript	1,602	7.6%
14	Distributed Systems	1,593	7.5%
15	Data Engineering	1,515	7.2%
16	Docker	1,505	7.1%
17	Azure	1,499	7.1%
18	Android	1,449	6.9%
19	SQL	1,449	6.9%
20	iOS	1,439	6.8%
21	React	1,405	6.6%
22	Autonomous Vehicles	1,345	6.4%
23	Deep Learning	1,284	6.1%
24	PyTorch	1,105	5.2%
25	Terraform	1,073	5.1%
26	Data Science	1,024	4.8%
27	Java	986	4.7%
28	Prompt Engineering	812	3.8%
29	Snowflake	758	3.6%
30	JavaScript	714	3.4%
31	Fine-tuning	688	3.3%
32	Robotics	679	3.2%
33	Node.js	674	3.2%
34	CUDA	671	3.2%
35	Foundation Models	657	3.1%
36	Rust	601	2.8%
37	PostgreSQL	592	2.8%
38	Reinforcement Learning	519	2.5%
39	Spark	517	2.4%
40	Computer Vision	506	2.4%
41	Kotlin	500	2.4%
42	RAG	462	2.2%
43	TensorFlow	430	2.0%
44	Embeddings	418	2.0%
45	C#	391	1.8%
46	JAX	333	1.6%
47	Kafka	328	1.6%
48	LangChain	327	1.5%
49	NLP	313	1.5%
50	Go	287	1.4%
51	Hugging Face	287	1.4%
52	MLOps	281	1.3%
53	Vector Databases	257	1.2%
54	Airflow	251	1.2%
55	Databricks	249	1.2%
56	Redis	244	1.2%
57	Swift	221	1.0%
58	Multimodal	203	1.0%
59	Transformers	203	1.0%
60	vLLM	202	1.0%
61	Diffusion Models	201	1.0%
62	Triton	193	0.9%
63	Mistral	171	0.8%
64	GraphQL	139	0.7%
65	pandas	135	0.6%
66	NumPy	95	0.4%
67	Scala	91	0.4%
68	scikit-learn	70	0.3%
69	MATLAB	66	0.3%
70	FastAPI	52	0.2%
71	Model Context Protocol	51	0.2%
72	Quantization	41	0.2%
73	TensorRT	38	0.2%
74	MongoDB	30	0.1%
75	Elasticsearch	29	0.1%
76	Tableau	26	0.1%
77	LangGraph	25	0.1%
78	MLflow	24	0.1%
79	Time Series	24	0.1%
80	BigQuery	23	0.1%
81	dbt	20	0.1%
82	Django	20	0.1%
83	LlamaIndex	19	0.1%
84	SageMaker	17	0.1%
85	Ruby	14	0.1%
86	Power BI	13	0.1%
87	ONNX	13	0.1%
88	CrewAI	13	0.1%
89	Flask	13	0.1%
90	Pinecone	12	0.1%
91	Vertex AI	10	0.0%
92	Llama	10	0.0%
93	Weaviate	10	0.0%

Share is the percent of active AI roles that mention the skill. Skills in fewer than 10 roles are omitted. Each skill links to its live openings.

About this data

How is skill demand measured?: We scan the title and description of every active AI job posting for a curated vocabulary of skills, then rank each skill by the number of distinct live roles that mention it. Share is that count divided by all active AI roles.
Which skills are counted?: Programming languages, ML frameworks, AI techniques (RAG, fine-tuning, agents), cloud and data tools, and current LLM tooling (LangGraph, LlamaIndex, vLLM, and more). Skills appearing in fewer than 10 live roles are omitted to avoid noise.
How current is it?: It reflects live listings and updates as new roles are posted and old ones close. Each skill links to its current openings.