AI and ML
Job Market Intelligence
Which AI skills are employers actually hiring for? Rankings for ML engineering, LLM development and AI operations — built from active job postings, not surveys. Pair with Hugging Face trending models to see where the industry is heading next.
Skills snapshot: 2026-07-17 · Updated weekly
Active AI listings
2,875
Live job postings
Top skill
Machine Learning
31% of AI listings
Top role
ML Engineer
501 open positions
HF trending now
all-MiniLM-L6-v2
sentence-transformers
Top AI and ML skills by demand
Snapshot: 2026-07-17
Machine Learning
LLMs / GenAI
Python
AI Agents
PyTorch
AWS
Stakeholder Mgmt
RAG
Azure
Fine-tuning
Deep Learning
TensorFlow
A/B Testing
Agile / Scrum
NLP
GCP
SQL
Statistical Analysis
Kubernetes
Spark
AI roles with the most open positions
Normalized from active job posting titles
ML Engineer
AI Engineer
Research Scientist
MLOps Engineer
Computer Vision Eng
Skill mix in top AI postings
Proportion of top 20 skills by category — from active AI and ML job postings.
Reading the AI hiring landscape
What the demand data means for engineers targeting AI and ML roles
Python is the entry point
Python appears in more AI and ML postings than any other skill — including ML frameworks. Teams building on PyTorch, TensorFlow or any GenAI stack list Python as a prerequisite. It is not a differentiator; it is the baseline assumption for every role on this list.
GenAI has split the skills list
LLM, RAG, transformers, embeddings and vector databases have moved into the top rankings. Many roles now split into two tracks: classical ML (scikit-learn, XGBoost, feature engineering) and GenAI (fine-tuning, retrieval-augmented generation, prompt pipelines). Identifying which track a posting targets tells you which skills to lead with.
MLOps is no longer optional
Cloud platform skills — AWS SageMaker, Azure ML, Vertex AI, Kubernetes, Docker and MLflow — appear alongside model skills in the majority of mid-level and senior postings. Engineers are now expected to deploy, monitor and iterate models without handing off to a dedicated infrastructure team.
PyTorch has overtaken TensorFlow
PyTorch now leads TensorFlow in most new-hire postings, driven by its dominance in research and adoption by major model providers. JAX is rising in high-performance computing roles. Hugging Face Transformers bridges both — it is the library most teams use regardless of their underlying framework preference.
Hugging Face trending models
A leading indicator of which model architectures and tasks are gaining developer attention
| # | Model | Trending |
|---|---|---|
| 1 | all-MiniLM-L6-v2 sentence-transformers | — |
| 2 | ms-marco-MiniLM-L6-v2 cross-encoder | — |
| 3 | bert-base-uncased google-bert | — |
| 4 | bge-small-en-v1.5 BAAI | — |
| 5 | paraphrase-multilingual-MiniLM-L12-v2 sentence-transformers | — |
| 6 | electra-base-discriminator | — |
| 7 | bge-m3 BAAI | — |
| 8 | all-mpnet-base-v2 sentence-transformers | — |
| 9 | mobilenetv3_small_100.lamb_in1k timm | — |
| 10 | Qwen3-0.6B Qwen | — |
| 11 | t5-small google-t5 | — |
| 12 | clip-vit-base-patch32 openai | — |
| 13 | xlm-roberta-base FacebookAI | — |
| 14 | Qwen3-8B Qwen | — |
| 15 | bge-reranker-v2-m3 BAAI | — |
| 16 | chronos-2 amazon | — |
| 17 | opt-125m | — |
| 18 | nomic-embed-text-v1.5 nomic-ai | — |
| 19 | bge-large-en-v1.5 BAAI | — |
| 20 | gemma-4-26B-A4B-it | — |
Top 20 by trending score · Data from Hugging Face Hub API · Score reflects recent download velocity and community engagement
How this intelligence is gathered
Methodology behind the rankings, refreshed weekly
Job listings
AI and ML job postings are scraped daily from company career pages and aggregated job boards. Each listing is deduplicated and tagged with a normalised role category. Skills are extracted from each description and stored as weekly snapshots — so the rankings reflect what employers are actively hiring for right now, not six months ago.
Skill rankings
Rankings reflect how frequently each skill appears in active job listings — a direct measure of employer demand, not survey opinions or social media popularity. Percentages shown are share of all current AI listings. A skill's rank can shift week-over-week as postings open and close. The category breakdown shows how demand is currently split across languages, frameworks, GenAI and cloud infrastructure.
HF models as forward signal
Hugging Face trending models show what the AI research and developer community is actively building with. The models at the top of the trending list today often become the architectures that production teams hire for over the next six to twelve months — making this table a useful early signal of skills demand ahead of the hiring curve.
Explore related intelligence
Skill Trends
Live rankings across all tech skills
Languages
Programming language demand index
Market Statistics
Snapshot metrics from the live index
Skill Spotlights
Deep dives on individual skills
AI Requirements Index — open dataset
Download the daily share of job listings requiring AI skills, by tier, category and seniority (CSV/JSON, CC BY 4.0)
Job data from active listings · Model data from Hugging Face Hub API · Updated weekly