Most rate cards are useless in the same way a menu with no prices is useless: they give you a number without the one thing that makes it mean anything. “$25/hour” tells you nothing if you don’t know whether that buys a junior who can wire up an API call or a senior who can tell you why your retrieval pipeline is quietly setting money on fire.
This is the reference page we point people to when they ask what an AI engineer costs. It is the full breakdown behind the summary figures in our guides on hiring AI engineers and AI DevOps engineers. The numbers are what we see in the market in 2026, cross-referenced against public salary data.
One thing up front: a rate card tells you the price, not the value. A senior who prevents one bad architecture decision pays for the difference over a junior many times over.
How to read this rate card
Three things before the tables, because they move the numbers more than seniority does.
These are gross monthly engineer rates, not your fully-loaded budget. On top of the figures below you add employer contributions, 13th-month accrual, equipment, and a partner or recruiting fee. In Vietnam that adds roughly 35-45%.
“AI engineer” is a band, not a job. The same title covers someone integrating a foundation-model API and someone building eval harnesses and retrieval infrastructure. Infrastructure-heavy AI DevOps and LLMOps work prices at a 15-25% premium over the figures below.
Market matters more than most people expect. The same seniority costs roughly 3-4x more in the US than in Vietnam. That gap is the reason offshore and embedded hiring exists.
Monthly rate card by seniority and market
All figures are monthly, in USD, for a full-time AI engineer in 2026. Ranges reflect the spread between competent and genuinely strong at each level.
Vietnam
| Seniority | Experience | Monthly (USD) |
|---|---|---|
| Junior | 0-2 years | $1,000 - $1,800 |
| Mid | 2-4 years | $1,900 - $3,000 |
| Senior | 5+ years | $3,300 - $5,000 |
| Lead / Principal | 8+ years | $5,000 - $7,500 |
Eastern Europe
| Seniority | Experience | Monthly (USD) |
|---|---|---|
| Junior | 0-2 years | $2,500 - $3,500 |
| Mid | 2-4 years | $3,500 - $5,000 |
| Senior | 5+ years | $4,200 - $6,500 |
| Lead / Principal | 8+ years | $6,500 - $8,500 |
Western Europe
| Seniority | Experience | Monthly (USD) |
|---|---|---|
| Junior | 0-2 years | $4,000 - $6,000 |
| Mid | 2-4 years | $6,000 - $8,500 |
| Senior | 5+ years | $7,000 - $10,500 |
| Lead / Principal | 8+ years | $10,500 - $14,000 |
United States
| Seniority | Experience | Monthly (USD) |
|---|---|---|
| Junior | 0-2 years | $6,000 - $9,000 |
| Mid | 2-4 years | $9,000 - $12,500 |
| Senior | 5+ years | $11,000 - $15,500 |
| Lead / Principal | 8+ years | $15,500 - $22,000 |
Hourly and contract rates
For project work, migrations, or fractional coverage, the same people price hourly. Contract rates carry a premium over the equivalent full-time cost because you are paying for flexibility and no long-term commitment.
| Market | Junior | Mid | Senior |
|---|---|---|---|
| Vietnam | $20 - $35 | $35 - $55 | $55 - $90 |
| Eastern Europe | $35 - $50 | $50 - $75 | $75 - $110 |
| Western Europe | $50 - $75 | $75 - $110 | $110 - $160 |
| United States | $75 - $120 | $120 - $175 | $175 - $275 |
How AI DevOps and LLMOps compare
Infrastructure-side AI roles — serving models, running retrieval pipelines, owning token cost and eval harnesses — sit 15-25% above the AI engineer bands above. A senior AI DevOps engineer in Vietnam lands around $3,500-$5,000/month; in the US, $12,000-$18,000.
What moves someone up a band
Seniority on this card is not years on a CV. Three things separate a mid-level AI engineer from a senior one, and none show up in a job title:
They have shipped to real users, not just a demo. The gap between “works in the notebook” and “works when 500 people hit it at once” is where most of the value lives.
They can reason about cost. A senior treats token spend as an owned metric, not a monthly surprise.
They can measure whether a change helped. Eval discipline is one of the clearest signals of seniority.
Why the cheapest number on this card is a trap
The spread across markets is the first thing everyone notices, and the wrong thing to optimise for.
A junior at $2,000/month who takes four months to ship what a senior ships in one, and ships it with a retrieval bug that quadruples your inference bill, is not cheaper. The salary line looks great. The total cost — delay, rework, and GPU spend — does not.
The market arbitrage is real: a senior Vietnamese AI engineer at $3,500/month genuinely costs a quarter of the US equivalent. But you capture it by hiring a strong senior offshore, not a cheap junior anywhere. The regional discount is worth taking. The seniority discount is not.
If you want a number for your stack rather than a range, talk to our AI engineering team. The first conversation is usually technical, not sales-led.