Forward-deployed engineer
Build AI systems that solve operational problems.
Neurai's FDE network is for developers who already code and want to deliver enterprise integrations and workflows. Demonstrate how you investigate, build, test and explain a solution, and get staffed on client projects.
The steps
Five short steps.
The times are estimates. Before each step your portal says what it is for, what is recorded and which tools you may use.
- 01
Application: about ten minutes
A CV, a GitHub profile or a LinkedIn profile (any one is enough), one system you built that runs in production, and a few specific questions about what the role needs (below). No salary history, birthday, photo or passport. Verify your email to continue. - 02
Neurai Field Lab: about 25 to 35 minutes
A reasoning exercise about a fictional operations problem, in four rounds with a tutorial first. It records the decisions and experiments you make inside the game, nothing else on your device. Your own notes and tools are allowed. When you finish, the practical task opens for you. - 03
Practical task: a 60-minute window
Debug and complete a small implementation on synthetic data, push it to a private GitHub repository, answer three business-data questions, and record an explanation of at most 5 minutes (or write the text equivalent). The task is unpaid. It is an assessment, not commissioned work, and its licence covers evaluation only. - 04
Review and admission
Independent tests run against your repository, and your work and your explanation are reviewed against the published criteria, together with your Field Lab record. When you pass, you are admitted at once and an email tells you that you passed; otherwise our team decides, with a reason you can read and appeal. - 05
LinkedIn, the final in-person interview, your employment and offer letter
Admission is a letter confirming you joined the Neurai FDE network. You then add your Neurai role to your LinkedIn profile from your portal: a member of Neurai’s staffing team needs to see it on your profile in order to reach out to you for the final in-person interview. After that interview, once you are selected for a project, you receive your employment and offer letter with the project and its terms.
Who should apply
What the role needs, said plainly.
Read this before you apply. The interview goes into each point.
The role is for you if
- Three years or more of paid work writing software that other people used.
- Python or TypeScript, and SQL: you can read a small codebase in one of them and query a relational database.
- You have connected software to a business system (an ERP, a CRM, an accounting or warehouse system) or to another company’s API, and it ran in production.
- You have built with a language model: a feature, an agent or a retrieval system that real users used.
- You have worked directly with the people who use a system you built: requirements, a demonstration, a go-live.
- Fluent English, written and spoken, and at least four hours of your working day inside US business hours.
- A GitHub account: the practical task is submitted through it.
It is not for you if
- You are learning to code or looking for a first developer job.
- Your experience is configuring a product rather than writing software.
- You want to train models or do research: this work is integration and delivery inside a customer’s systems.
- You would rather write the code and leave the customer to someone else.
- A contract or an exclusivity forbids you to take on this work.
Questions people ask
Before you apply.
- What is the practical task, and what is it not?
- A 60-minute window, started when you choose: debug and complete a small implementation on synthetic data, push it to a private GitHub repository, answer three business-data questions and record an explanation of at most 5 minutes, or write it. The task is unpaid. It is an assessment, not commissioned work, and its licence covers evaluation only. What you produce runs only in an isolated environment for evaluation and is never used commercially.
- Which hours, and from where?
- At least four hours of your working day inside US business hours, and fluent English, written and spoken, for work with the people who use the system.
- How are you assessed, and what are your rights?
- Your Field Lab, your practical task and your explanation are assessed against the published criteria. Nobody is turned down automatically: any decision that holds you back is taken by our team, with a reason you can read and appeal. You can ask our team to look at any step again, for an accessible equivalent of any step, or for an adjustment: the help page says how, and the privacy notice has the detail.
- Is there a fee, and what do we never ask for?
- No fee of any kind, ever. The form asks for no salary history, birthday, photo or passport. A face camera is never required: a screen recording with your voice is enough.
What we assess, and what we never do
The steps show
How you work, on real tasks
- That you can read a small Python or TypeScript codebase, find what is wrong and finish it.
- That you understand HTTP APIs, relational data and version control with Git and GitHub.
- That you can explain a decision plainly, with the records it rests on.
We never
Judge anything but your work
- Infer intelligence, personality, honesty or health from a game, a video or a CV.
- Judge accent, appearance, speed, follower counts or prestige. Screen plus narration is enough; a face camera is never required.
- Turn anyone down automatically. A decision that holds you back is taken by our team, with a reason you can read.
Admission is the start.
After admission you add your Neurai role to LinkedIn so that Neurai’s staffing team can reach out to you; the final in-person interview follows and, once you are selected for a project, your employment and offer letter with its dates and its terms.