MLOps and AI Platform Engineer Jobs in 2026: 150 Roles, Almost No Juniors
Every ML model that ships lives on infrastructure someone had to build. The team that built it is now one of the most in-demand — and hardest to break into — in AI hiring.
HireIndex’s current index of 1,541 open AI/ML roles across 655 companies (as of June 8, 2026) includes 150 roles explicitly tagged as MLOps Engineer, AI Platform, or AI Infrastructure. That’s 9.7% of the entire index — roughly the same footprint as AI Research — for a category that rarely gets coverage relative to the flashier engineering titles.
One number stands out above the rest: of the 150 roles in this category, just one is entry-level.
How many MLOps and AI Platform jobs are available right now?
Here’s the current breakdown across the three infrastructure-layer skill categories:
| Skill | Open Roles | Share of Index |
|---|---|---|
| AI Platform | 76 | 4.9% |
| MLOps Engineer | 61 | 4.0% |
| AI Infrastructure | 33 | 2.1% |
| Combined | 150 | 9.7% |
AI Platform leads narrowly. The three categories overlap conceptually — all involve building and operating the systems that ML engineers use to train and deploy models — but employers tend to use them for different seniority levels and organizational contexts. AI Platform skews toward product-minded engineering inside large companies building internal developer tools. MLOps Engineer titles concentrate at mid-sized companies and consultancies deploying ML at scale. AI Infrastructure tends to appear at frontier labs (OpenAI, Scale AI) building hardware-adjacent software for compute-intensive workloads.
For context: LLM Engineer has 37 roles in the same index — the title that generated the most candidate interest when it emerged. MLOps and AI Platform combined are four times that volume.
Which companies are hiring for AI infrastructure roles?
Capital One leads the category with 15 open roles — more than any AI-first company, including organizations that build infrastructure as their core product.
| Company | Open Roles |
|---|---|
| Capital One | 15 |
| Databricks | 7 |
| NavitasPartners | 7 |
| OpenAI | 5 |
| Scale AI | 4 |
| Spotify | 4 |
| Figure AI | 4 |
| 3 | |
| Deepgram | 3 |
Capital One’s titles tell the story precisely: Sr. Lead AI Engineer (Inference Optimization, FM Hosting, AI Platform), Lead AI Engineer (Gen AI Platform, Agentic AI & LLM Infrastructure & Orchestration), Distinguished AI Engineer (Agentic AI Platform). This is not generic cloud infrastructure work. It’s the engineering discipline of making large language models run reliably and cheaply inside a bank serving 75 million customers — where latency targets and cost-per-inference are real business constraints, not benchmarks.
Databricks is the logical second — the platform-of-platforms, where data scientists and ML engineers do their work. Seven open roles span architecture, engineering, and product: Staff Software Engineer — AI Platform, Staff Product Manager, AI Platform, Data & AI Platform Architect (Professional Services). Their AI Platform hiring operates on two tracks simultaneously: building the product and deploying it for enterprise customers.
OpenAI’s five infrastructure roles reveal where Stargate compute investment flows in practice: Manufacturing Test Engineer, AI Compute Infrastructure, Supply Chain Program Manager — AI Infrastructure, Software Engineer, Monetization ML Infrastructure. OpenAI is building the physical and software infrastructure for next-generation compute, not just the models running on it. The supply chain and manufacturing titles in particular have no equivalent anywhere else in the index.
Where are MLOps and AI Platform jobs located?
This is the most remote-accessible major category in the current index: 73% of all MLOps/AI Platform/Infrastructure roles list as remote (109 of 150).
| Location | Open Roles |
|---|---|
| Remote | 109 |
| London | 16 |
| San Francisco | 8 |
| New York | 5 |
| Toronto | 4 |
| Sydney | 3 |
For comparison, Machine Learning Engineer roles run about 63% remote across the full index. Infrastructure roles skew further distributed — reflecting both the nature of the work (services that run everywhere, owned by teams that don’t need to be co-located to operate them) and the company profiles driving hiring (enterprise organizations with distributed engineering organizations).
London leads in-person with 16 roles, split between Capital One’s UK tech operation and a cluster of European financial services firms. San Francisco’s 8 roles concentrate at OpenAI and Scale AI. The overall in-person market for these titles is thin — if you need an office, your city options narrow considerably compared to ML engineering or data science.
Why are there almost no junior-level MLOps roles?
One. One junior-level role across the 150 in this combined category.
The full seniority breakdown:
| Level | Count | Share |
|---|---|---|
| Mid-level | 87 | 58% |
| Senior | 41 | 27% |
| Senior-heavy (Staff / Principal) | 20 | 13% |
| Leadership | 1 | 1% |
| Junior | 1 | 1% |
Senior+ concentration sits at 41% — higher than LLM Engineer (38%) and significantly higher than Machine Learning Engineer. This is one of the most experience-weighted categories in AI hiring.
The reason isn’t gatekeeping for its own sake. MLOps and AI Platform engineers own production infrastructure: training clusters, inference APIs, monitoring pipelines, and feature stores that power everything else. A mistake doesn’t cause a bad eval result — it causes an outage that reaches real users at scale. Production ownership requires the pattern-recognition that comes from having shipped and broken things before. That’s what the seniority concentration reflects.
The practical entry path: nearly every company in this category is also hiring Machine Learning Engineer roles, which do carry a meaningful junior surface across 313 indexed openings. Spend 2-3 years in applied ML work — building, shipping, and operating models — and you accumulate the production ML intuition that MLOps and AI Platform interviews are designed to test. The infrastructure layer is the second job, not the first.
What does this work actually look like day-to-day?
Capital One’s open titles are the most detailed window: Agentic AI Platform, LLM Inference, GenAI Platform Services, MLX (ML Experimentation). The work is the plumbing of modern enterprise AI: orchestrating LLM calls across services, managing inference latency and cost at scale, building internal platforms that dozens of ML teams depend on without thinking about.
Databricks titles add the enterprise view: AI Platform Architect (Professional Services) means implementing the platform for other companies’ data teams — you’re the expert at the frontier of what the product can do, working on the hardest edge cases in a client’s environment. Scale AI’s infrastructure roles focus on model serving and training platforms — the compute infrastructure behind large-scale data annotation and RLHF work at the highest volumes in the industry.
The common thread across all three profiles: these roles are less about building models than about making sure models built by others actually run — reliably, cheaply, and at a cost structure that makes the business case work. As inference from frontier models becomes a line item that CFOs scrutinize, that last part is its own specialized discipline. The engineers who understand both ML systems and infrastructure cost optimization are in a category with very few credentialed candidates.
What this means for your job search
Target the real employers. Capital One, Databricks, and OpenAI account for 27 of the 150 roles — 18% of the category. If you’re targeting AI Platform or MLOps specifically, these three companies represent a disproportionate share of the actual openings. Capital One in particular is running the largest AI Platform hiring program in the current index — a fact that surprises most candidates who assume the market clusters at AI-first companies.
Remote is genuinely real here. 73% remote in this category is not an artifact of vague ATS fields — it reflects how infrastructure teams actually work at companies like Databricks, Scale AI, and Reddit. If location flexibility is a priority, this is one of the stronger categories to target.
Mid-level is the real entry surface. With essentially no junior seats, the addressable market for early-career candidates is near zero. If you’re 1-2 years in and interested in infrastructure work, focus first on Machine Learning Engineer roles — 313 openings with a real entry-level slice. The infrastructure path opens after you’ve operated ML systems in production.
Search wider than “MLOps.” The category is fragmented across titles. “AI Platform Engineer,” “ML Infrastructure Engineer,” “ML Systems Engineer,” “GenAI Platform Engineer,” and “AI Infrastructure Engineer” all describe similar work and appear across the same companies. Filtering by exact title misses the majority of this market.
Browse current AI Platform, MLOps Engineer, and Machine Learning Engineer roles in the full index. Data updated every Monday from direct ATS and aggregator sources.
HireIndex tracks 1,541 open AI/ML roles across 655 companies as of June 8, 2026.