Forward deployed engineersshipping your backlog on Steer.
A pod of engineers embeds with your team, steers AI agents in parallel and ships every pull request with evidence.
We build the pod for your work, for example four engineers, with the machine learning, software, DevOps and MLOps talent you need, for a monthly fee set by the pod's size. They work with Steer by Pento, our platform, in your GitHub, under the merge policy you set.
What a forward deployed team does
- Embedded with your teamIn your Slack, your GitHub and your meetings
- Your backlog, as goalsAcceptance criteria before any code is written
- Agents in parallel, steeredSeveral sessions per engineer, each in its own sandbox
- Evidence on every pull requestThe tests that ran, the findings and screenshots
- Your merge policyMerges follow the autonomy level you choose
- Your GitHub, your codeIn your repositories from the first commit
What is a forward deployed engineer?
A forward deployed engineer is a software engineer who works embedded with a customer's team, close to the people who have the problem, and owns the outcome rather than a queue of tickets. Companies like Palantir made the role well known. At Pento, the role is a pod we build for you: a team of engineers who embed with yours, for example four of them, with the machine learning, software, DevOps and MLOps talent your work needs. The monthly fee depends on the pod's size. They work with Steer by Pento, our platform, steer AI coding agents in parallel and ship every pull request with evidence.
01Goals
02Build
03Review
04Merge
05Numbers
You
Decide the order of your backlog
Choose the autonomy level, approve what it gates
See the week next to last week's, from your seat
Forward deployed engineers
Turn each item into a goal with acceptance criteria
Steer a session when it starts to drift
Judge every change against its criteria
Steer + AI agents
Ask the person who knows when a rule is missing
Several sessions per engineer, each in its own sandbox
Attach the tests, findings and screenshots
Count sessions, merges, lead time and interventions
Review + checks
An independent review model checks every change
Low-risk changes merge on a green review and CI
- 01Goals
- YouDecide the order of your backlog
- Forward deployed engineersTurn each item into a goal with acceptance criteria
- Steer + AI agentsAsk the person who knows when a rule is missing
- 02Build
- Forward deployed engineersSteer a session when it starts to drift
- Steer + AI agentsSeveral sessions per engineer, each in its own sandbox
- 03Review
- Forward deployed engineersJudge every change against its criteria
- Steer + AI agentsAttach the tests, findings and screenshots
- Review + checksAn independent review model checks every change
- 04Merge
- YouChoose the autonomy level, approve what it gates
- Review + checksLow-risk changes merge on a green review and CI
- 05Numbers
- YouSee the week next to last week's, from your seat
- Steer + AI agentsCount sessions, merges, lead time and interventions
The rhythm repeats every week, in the order you set.
The shape
Monthly, embedded, and often the team that built it.
One pod and one monthly fee set by its size, written down before the team starts. No hourly billing.
- One quote per pod
Monthly, set by pod size
Billed monthly, per pod. We quote after the free audit: the number of engineers we allocate sets the monthly fee, written down before the team starts.
- Embedded with your team
A pod of four, or the talent you need
We staff the pod with the machine learning, software, DevOps and MLOps engineers your work calls for, sized during the audit. Every engineer on it writes specs, steers agents and reviews evidence, and the people you meet are the people who ship.
- Often
After a fixed-price project
When a fixed-price project ships, the same engineers who built it can stay on and keep working through your backlog.
Compared
Forward deployed team vs staff augmentation vs consultants vs an AI coding tool.
Each one is the right call for someone. The differences are who runs the agents, what you can see while the work happens, and how you know a change is done.
| Compared on | Forward deployed team | Staff augmentation | Consultants | An AI coding tool |
|---|---|---|---|---|
| What you buy | Forward deployed teamA pod of forward deployed engineers working your backlog, billed monthly by pod size | Staff augmentationEngineers' time, added to your team and your process | ConsultantsAdvice and a plan, sometimes a build, usually billed by the hour | An AI coding toolSeats or usage, operated by your own engineers |
| Who runs the agents | Forward deployed teamOur engineers, on Steer, several sessions at a time | Staff augmentationWhoever you assign, with the tools you choose | ConsultantsDepends on the firm | An AI coding toolYour engineers, alongside the rest of their work |
| What you can see | Forward deployed teamSpecs, sessions, evidence and weekly numbers, live in your workspace seat | Staff augmentationYour own tracker and standups | ConsultantsStatus reports and decks | An AI coding toolUsage dashboards |
| How you know it's done | Forward deployed teamEvidence on every pull request, checked against the goal's criteria | Staff augmentationYour code review process | ConsultantsWhatever the statement of work defines | An AI coding toolYour code review process |
| Best when | Forward deployed teamA backlog needs experienced people and more throughput than hiring can add | Staff augmentationYou want specific skills inside your team, under your process | ConsultantsYou need an outside view or a strategy before you build | An AI coding toolYour team has time to operate agents and review what they write |
Want engineers inside your own process, without Steer? That is classic staff augmentation, and we do that too: see AI team augmentation.
How a week runs
A week with a forward deployed team.
The rhythm repeats every week: questions go to the person who knows, sessions get steered while they run, and the numbers land in your seat.
Needs you questions
When a goal is missing a rule, Steer asks the person who knows, in Slack or the Needs you inbox, with a deadline and a default. If nobody answers in time, the default applies and the work keeps moving.
Steering sessions
Our engineers watch the sessions they run and steer the ones that drift. A steer message lands at the next five-minute slice, so a wrong turn costs minutes instead of a rejected pull request.
Reviews and merges
Every pull request gets an independent review and a forward deployed engineer's verdict. Merges follow the autonomy level you choose, from Propose, where a person decides every merge, to Gated, where low-risk changes merge on a green review and CI. Whatever the policy gates waits for your reviewer.
Weekly numbers
Your Viewer seat shows spend, sessions, pull requests merged, lead time and how often a person stepped in, each next to last week's.
FAQ
What teams ask before hiring forward deployed engineers
Not covered here? Bring the question to a discovery call. An engineer reads every request.
Book a discovery callA forward deployed engineer is a software engineer who works embedded with a customer's team, close to the people who have the problem, and owns the outcome rather than a queue of tickets. Companies like Palantir made the role well known. At Pento, the role is a pod we build for you: a team of engineers who embed with yours, for example four of them, with the machine learning, software, DevOps and MLOps talent your work needs. The monthly fee depends on the pod's size. They work with Steer by Pento, our platform, steer AI coding agents in parallel and ship every pull request with evidence.
Staff augmentation adds engineers to your team, your process and your tools, and you direct their work. Pento's forward deployed engineers also embed with your team, but they bring Steer: each backlog item becomes a goal with acceptance criteria, agents build in parallel, and every pull request ships with evidence you can check. If you want engineers inside your own process without Steer, that is our AI team augmentation service.
Monthly, per pod. We quote after the free audit, once we have seen your backlog and the systems around it. The number of engineers we allocate sets the monthly fee, and it is written down before the team starts. We don't publish rates, because the pod and the work set the quote.
We size the pod to your work during the audit. Four engineers is one example. We staff it with whatever talent you need across machine learning, software, DevOps and MLOps, and every engineer on it writes specs, steers agents and reviews evidence.
Our engineers work from Latin America, and their working hours overlap with US time zones, so questions, steering and reviews happen during your working day.
No. Our engineers operate Steer. Your team gets a seat in the same workspace from day one: a Viewer seat shows progress, spend, speed and the full record, and a Reviewer seat also answers questions and decides merges. Questions can also reach your people in Slack.
With the free audit: a written report on your backlog, the systems you run and the code behind them. From there you can start with the team. If there is a clearly scoped piece of work to do first, such as an integration, you can start with it as a fixed-price project and keep the same engineers on as a forward deployed team when it ships.
You do. The code lives in your GitHub repositories from the first commit, on branches and pull requests, and the full history stays with you if we part ways. Each run uses a short-lived token that reaches only your project's repositories.
Contact us
Put a forward deployed team on your backlog
Start with the free audit. We read your backlog and the systems around it, then quote a monthly fee for a pod of forward deployed engineers, sized to your work. An engineer reads every request.