MLOps: mercado
MLOps is the engineering of the machine-learning model lifecycle: reproducible training, model registry, production rollout, quality and data-drift monitoring. We track 2 195 open positions listing MLOps, 2 059 of them fresh. LLMs widened the role: classic pipelines now come with LLM application evaluation, GPU model serving and inference cost management. Most of the demand comes from ML/AI, AI Engineer and Data roles. It most often appears next to Python, AWS and LLM. 34% of the postings allow fully remote work. Below is measured MLOps demand.
O mercado de MLOps tem agora 2.195 vagas abertas, 2.059 delas observadas recentemente. Median salary $10.507/mo. Observed candidate pool — 356.
34% of MLOps postings allow fully remote work; the share counts only postings that state the work format.
⚠ salary known for 176 of 2195 jobs; remote share from 1759 with a stated format; 2059 counted as fresh observations; trend and hiring difficulty are not shown
Demanda e oferta observada
| Demanda aberta | 2.195 |
| Oferta observada | 356 |
⚠ candidates matching a vacancy are not counted yet: demand and the observed pool are shown
Distribuição de salários
Um em cada dez ganha menos de $4.250, um em cada dez mais de $17.916. Metade do mercado fica entre $6.250 e $15.416. Amostra: 176.
The range histogram for MLOps postings uses only postings with an explicit compensation range; the median is computed over the same sample.
Geografia da demanda
| país | vagas |
|---|---|
| US | 441 |
| PL | 306 |
| GB | 177 |
| IN | 143 |
| DE | 132 |
| FR | 61 |
| CA | 57 |
| RU | 45 |
| SG | 42 |
| BR | 40 |
The top country for MLOps demand is the US (441 positions).
⚠ job counts only: salary by country is not published
Usada junto com
MLOps most often appears next to Python, AWS and LLM.
Cargos que a usam
| cargo | vagas |
|---|---|
| ML/AI | 788 |
| AI Engineer | 448 |
| Data | 282 |
| DevOps | 219 |
| Architect | 150 |
| Platform Engineer | 57 |
| Fullstack | 46 |
| Backend | 36 |
| Product | 26 |
| Security | 19 |
⚠ the `Other` and `IT` buckets are excluded as non-roles
Demanda por nível
| nível | vagas |
|---|---|
| senior | 754 |
| principal | 204 |
| lead | 163 |
| middle | 129 |
| junior | 35 |
Senior-and-above is about 87% of 1 285 graded MLOps postings. MLOps requires experience in both ML and infrastructure — junior openings are almost absent.
⚠ demand side only: the grade of the observed pool is unknown for most of it
Empregadores
A demanda está distribuída entre 1.189 empregador. O maior responde por 2.3%, os dez primeiros por 10.2%; os 89.8% restantes são cauda longa.
| proporção | valor |
|---|---|
| top 1 | 2.3% |
| top 3 | 5.0% |
| top 10 | 10.2% |
| cauda longa | 89.8% |
⚠ names are not shown: staffing agencies and end employers are not yet told apart by the classifier
Onde o Zorky vê este mercado
Observado em 100 fontes; a maior responde por 22.2%: este mercado não depende de um único canal.
MLOps postings arrive mostly from ATS & career sites and job boards & aggregators.
Vagas recentes
- Manager, Desjardins Artificial Intelligence Centre of Excellence
- AI Specialist · PL
- Machine Learning Operations (MLOps) Engineer · US
- Senior AI Product Manager · MG
- Machine Learning Engineer · GB
- Founding AI Engineer · GB
- Machine Learning Operations-Engineer II · US
- AWS Solution Architect · AE
Mercados adjacentes
⚠ adjacent markets for comparison are not defined yet
Como é medido
- Vaga
- an open job that cleared the quality gate and lists at least two technologies
- Candidato observado
- a candidate whose stack contains this technology; an aggregate — not a single record leaves the perimeter
- Candidato compatível
- not counted yet
- Janela
- jobs open at the moment the snapshot was built
Sobre os dados
- Alguns detalhamentos estão ocultos: a cobertura de dados deles ainda não é suficiente.
- As estatísticas só aparecem onde a amostra passa por um controle de qualidade.
- A ausência de um bloco não significa valor zero.
Detalhamentos ocultos no momento: 7.
Dados de 2026-10-26
Perguntas frequentes
Common questions about the MLOps market — short answers from the same data slice.
Which tools appear in MLOps postings?
MLflow or Weights & Biases for experiments, orchestrators (Airflow, Kubeflow, Prefect), Kubernetes and Docker, model serving (Triton, vLLM, KServe), cloud ML platforms (SageMaker, Vertex AI, Azure ML), monitoring.
How does MLOps differ from DevOps?
Besides code, data and models are versioned; quality is checked with model metrics, not just tests; production degrades from data drift even without code changes.
How big is MLOps demand?
The slice tracks 2 195 open MLOps positions; demand grows as teams move from experimenting with models to operating them.