You have probably seen the title by now. Forward Deployed Engineer. It sounds like a military posting, it pays like a senior software job, and if you ask three people what it means you get three answers.
So we went and counted the actual jobs.
On October 6, 2026 there were 275 open forward deployed engineer roles across the tech companies we track (distinct roles, so one role posted in ten cities counts once). That is 113 companies, and about 2.4% of every open tech role on the board. Small, but growing: among companies we have tracked since before mid-May, forward deployed titles were 1.7% of the roles opened between early June and early August, and 2.6% of the roles opened since early September. That is a 50% jump in share in a few months.
What the job is
The short version: a software engineer who sits with the customer instead of with the product team. You take the platform your company sells and make it work inside one specific customer's mess (their data, their cloud, their security rules, their half-finished internal tools), and you write real code to do it. Then you carry what you learned back to the people building the product.
The title is usually credited to Palantir, which used it for years before it spread. What changed in the last couple of years is that AI companies picked it up, because a model on its own does not do anything for an enterprise until somebody wires it into that enterprise.
The stack on these roles tells the same story:
Python in 166 of the 275 (60%)
AWS in 87, GCP in 74, Azure in 68
TypeScript in 71
Kubernetes in 63
RAG in 29, LLMs in 28
Look at the three clouds. They sit close to each other. That fits the work: you do not get to pick the cloud, the customer already did.
Who is hiring them
| Company | FDE roles |
|---|---|
| Databricks | 31 |
| OpenAI | 22 |
| Telnyx | 14 |
| Cohere | 14 |
| Scale AI | 8 |
| Notion | 6 |
| Cloudflare | 5 |
| Handshake | 5 |
| Invisible Technologies | 5 |
| Modus Create | 5 |
| Nebius | 5 |
You might expect this to be an AI-lab title. It mostly is not. By our count, roles at companies whose product is AI make up about 37% of forward deployed openings (101 roles at 32 companies). The other 63% sit at 81 companies that sell data platforms, communications APIs, productivity software, networks and consulting. Count Databricks as an AI company and it is closer to half and half. Either way, this is not a job you can only get at three labs in San Francisco.
The long tail is long: 113 companies, and most have only one to three of these open. Three of the titles even say "Founding Forward Deployed Engineer", which is about as early as a hire gets.
What it pays
129 of the 275 roles publish a yearly US dollar band. The median floor is about $181,000 and the median ceiling about $250,000.
That sits a little above the median Python role on the board (a $168,650 floor) and below machine learning titles ($200,000). It is not the jackpot some posts about this job make it sound like. It is a well paid engineering job with customers attached.
Who gets in
Here is the encouraging part. Half of forward deployed roles are mid-level (142), 130 are senior, and only 3 are junior. Compare that with machine learning engineering, where 69% of roles are senior, or engineering management, where it is 85%.
So if you are a few years in, can code, and are good with people who are not engineers, this is one of the more open doors in the market right now. If you are fresh out of school, it mostly is not (three junior roles across 113 companies).
Where you work from: 46% hybrid, 31% remote within a country or region (mostly the US), 20% onsite, and 8 roles labelled remote from anywhere. The job is called "forward deployed" for a reason, so expect some of it to happen near customers.
If you are thinking about it
Search the title, then search "solutions engineer" and "deployment engineer" too. Companies are still settling on a name, and similar work turns up under those titles.
Get comfortable in more than one cloud. The roles do not agree on which one, because the customers do not.
Read how much of the job is code. Postings with the same title can describe quite different days, so check the responsibilities, not just the name.
If you want to work at an AI company but not on models, this is one of the more open ways in.
Every role on remoet.dev/jobs carries its stack, seniority and remote policy, and an agent can search the board without an account.
What we could not measure
We matched "forward deployed" in the job title, so we miss similar jobs under other names (point 1 above), and we cannot tell from a title how much of a given role is code. The AI or not-AI split is our own judgement call on each company's main product. Salaries cover only roles that publish a band. The growth number compares the share of newly opened roles in two windows at companies we tracked across both, so it measures where hiring is moving, not the size of the market. And it is all one day, October 6, 2026.
If you are a forward deployed engineer and this does not match your job at all, we would really like to hear what does. Come tell us in Discord.
