Breaking Into AI in 2026: Only 41 Entry-Level Roles Exist Across 1,552 Openings
Every post about the AI talent shortage assumes the problem is a lack of qualified candidates. The hiring data tells a different story — at least for anyone early in their career.
Of the 1,552 open AI/ML roles HireIndex tracks across 654 companies as of June 15, 2026, exactly 41 are classified as junior or entry-level. That’s 2.6% of the market. And when you look at what those 41 roles actually are, the number gets worse.
How many AI jobs are genuinely entry-level right now?
41 is the raw count. Here’s the breakdown by what those roles actually represent:
- Internships: 18 of the 41 are explicitly titled as internships — at Cresta (5), Meta (2), Databricks (1), Scale AI (1), Cohere (1), Waabi (1), Autodesk (1), SAP (1), and others. These are time-limited positions, not permanent headcount. Strip them out and you have 23 “junior” roles.
- Staffing agencies: SynergisticIT accounts for 6 of the 41 with titles like “Junior Java Developer – React.js & Spring AI Focus” and “Junior Machine Learning Engineer-remote/AI/Data scientist.” These are recruiter postings, not direct hires at the companies doing the AI work. Strip those and you’re at roughly 17.
- Genuine junior full-time roles: The remaining ~17 are spread across companies like Easygo (2 junior MLOps roles), Bending Spoons (1 graduate AI software engineer), Infinitive (1 junior ML engineer in New York), and a thin tail of single postings at firms most candidates have never heard of.
Seventeen real entry-level AI jobs across the entire tracked market. That’s the actual number.
Which AI job categories have the most — and fewest — junior openings?
The seniority gap varies dramatically by role type. Some categories are structurally closed to early-career candidates. Others have at least a crack in the door.
| Category | Total Roles | Junior Roles | Junior % | Senior+ % |
|---|---|---|---|---|
| Machine Learning Engineer | 301 | 9 | 3.0% | 45.5% |
| AI Research | 124 | 6 | 4.8% | 27.4% |
| Data Scientist | 202 | 6 | 3.0% | 39.1% |
| MLOps Engineer | 65 | 2 | 3.1% | 35.4% |
| AI Platform | 77 | 2 | 2.6% | 59.7% |
| AI Software Engineer | 93 | 2 | 2.2% | 58.1% |
| LLM Engineer | 26 | 1 | 3.8% | 26.9% |
| Applied AI | 117 | 1 | 0.9% | 27.4% |
| AI Infrastructure | 31 | 0 | 0.0% | 41.9% |
| AI Product Manager | 31 | 0 | 0.0% | 51.6% |
| AI Solutions | 26 | 0 | 0.0% | 50.0% |
| Data Engineer | 14 | 0 | 0.0% | 64.3% |
Four entire job categories — AI Infrastructure, AI Product Manager, AI Solutions, and Data Engineer — have zero junior roles. Not a small number. Zero.
AI Research has the highest junior share at 4.8%, but those 6 roles are almost all PhD research internships at places like Meta and GenBio AI. If you don’t have a PhD, that 4.8% isn’t your 4.8%.
Where do junior AI roles concentrate geographically?
Remote dominates: 34 of the 41 junior roles (83%) are listed as remote. Only 7 have a specific city attached.
| City | Junior Roles | Total AI/ML Roles |
|---|---|---|
| Remote | 34 | 978 |
| London | 2 | 171 |
| New York | 1 | 76 |
| San Francisco | 1 | 123 |
| Toronto | 1 | 63 |
| Berlin | 1 | 9 |
| Los Angeles | 1 | 5 |
San Francisco — the city with the densest concentration of AI companies — has one junior role in the current index. London, the largest non-remote AI hiring market with 171 roles, has two. If you’re early-career and targeting a specific city for in-person AI work, the math is brutal: you’re competing for a handful of seats in markets with hundreds of mid-to-senior openings.
The remote concentration isn’t entirely good news, either. Many of those 34 remote junior listings are the internships and staffing agency postings described above. The genuine remote junior opportunities number closer to 10-12.
Why is the AI job market so hostile to entry-level candidates?
Three structural reasons — none of which are going away soon.
The work requires expensive mistakes to have already happened. MLOps, AI Platform, and AI Infrastructure roles — the categories with the highest senior+ concentration (42-64%) — exist specifically to prevent production failures. Companies don’t hire junior engineers to manage training clusters or inference pipelines serving millions of users. The pattern recognition required comes from having broken things before. This isn’t gatekeeping for its own sake; it’s risk management.
The companies doing the most AI hiring are mature organizations. Capital One (69 roles), Databricks (58), Amazon (37), and Spotify (27) are not building AI capabilities for the first time. They’re expanding established teams that already have junior-to-senior pipelines filled internally. External junior hiring at these organizations is rare because their internal talent development already produces the pipeline. The opening is at the senior end, where they need specialized experience they can’t grow fast enough.
AI labs are hiring researchers, not trainees. OpenAI (94 roles), Mistral AI (64), Scale AI (42), and Cohere (14) concentrate their non-senior hiring at the PhD intern level. The business model of a frontier AI lab doesn’t have a natural entry point for candidates without either a PhD or significant production ML experience. Their junior hiring is effectively an academic recruiting pipeline, not an open job market.
Which companies actually hire junior AI talent?
Cresta stands out: 5 internship roles out of 25 total AI/ML openings — a 20% junior share. That’s roughly 8x the market average. Cresta builds AI copilots for customer-facing teams, and their internship pipeline spans ML engineering, data science, product design, and forward-deployed engineering. For a mid-stage company with genuine AI work, that’s an unusually deep entry-level commitment.
Beyond Cresta, the real junior hiring happens at companies most candidates aren’t targeting:
- Easygo (2 junior MLOps roles) — an online gaming company, not an AI-first brand
- Bending Spoons (1 graduate AI software engineer) — European app company
- Infinitive (1 junior ML engineer, New York) — consulting firm
- Cohere (1 ML intern/co-op) — one of few frontier-adjacent companies with genuine intern volume
The pattern: companies that are building AI capabilities inside non-AI businesses, rather than companies that are AI-native. The former need to grow talent because they can’t compete on compensation for experienced AI engineers. The latter can afford to pay for experience and do.
What should early-career candidates actually do?
Target the mid-level floor, not the junior ceiling. The 946 mid-level roles (61% of the index) are the real addressable market. Many of these list “2-4 years of experience” — which is achievable through adjacent work in software engineering, data analysis, or research, combined with demonstrable ML project work. The bar is lower than it reads: companies posting mid-level AI roles are frequently open to strong candidates with 1-2 years plus clear evidence of relevant skills.
Go where the volume is. Machine Learning Engineer at 301 total roles has the largest absolute junior surface (9 roles) and the broadest mid-level tier. Data Scientist at 202 roles is the second-largest category and carries less competition per seat than the flashier titles. These two categories combined represent a third of the entire market.
Target non-AI-native companies. Companies building AI inside existing businesses — financial services firms, healthtech, enterprise software, gaming companies — are more likely to hire and develop junior talent than AI labs or AI-first startups that can afford to recruit experienced engineers exclusively.
Remote is your best shot. 83% of junior roles are remote. Don’t narrow your search to a specific city unless you have a compelling reason — the in-person junior market barely exists.
Internships are the pipeline, not the consolation prize. Cresta, Cohere, Meta, Databricks, and Waabi are all running AI internship programs. Converting an internship into a full-time offer at a company already doing the work you want to do is a stronger career move than landing a junior title at a staffing agency.
Browse the full index by role: Machine Learning Engineer, Data Scientist, or LLM Engineer. Filter by Remote to see the broadest entry-level surface. Data updated every Monday from direct ATS and aggregator sources.
HireIndex tracks 1,552 open AI/ML roles across 654 companies as of June 15, 2026.