# Why Eudia deploys lawyers alongside its engineers

> Its forward deployed lawyers need four years of practice and sit in Product, not Engineering, a sign of who defines 'correct' in regulated industries.

Original: https://fdetimes.net/en/analysis/eudia-forward-deployed-lawyers-regulated-industries/

Eudia's careers page lists two roles that both carry the words "Forward Deployed", but they sit in different sections. Forward Deployed Engineer is under Engineering. Forward Deployed Lawyer is under Product, and it requires at least four years of legal practice and the ability to handle matters independently, without supervision.

Eudia is a legal AI start-up founded in 2023. A job posting for its government practice, published on 30 June 2026, places the FDL on site at clients within the DoW, and Eudia states plainly that this is not a staff attorney position.

Engineers who work as FDEs, or want to, should pay attention to this detail. When a company hires both engineers and lawyers for client-facing work, the engineer is no longer the only person bringing knowledge back from the field.

Engineers make sure the system runs. Domain experts decide what counts as "correct". Understanding how these two roles complement each other will determine whether you are seen as a problem-solver or merely someone who wires up APIs.

## Why does an AI company need real lawyers?

The answer lies in the business model. Eudia calls its platform "Augmented Intelligence": combining AI with human expertise to help legal professionals work faster and better. The company has also acquired alternative legal service providers (ALSPs) such as Johnson Hana and Out-House.

It has gone further still with Eudia Counsel. According to chief executive Omar Haroun, this is the first "AI-native" law firm, operating under Arizona's ABS programme.

The firm takes on due diligence for M&A deals and contract work that the law requires a licensed lawyer to handle. In other words, there are deliverables that software is not permitted to sign off, however accurate it is.

Eudia Counsel also charges fixed fees rather than billing by the hour. When the price is set in advance, every hour of lawyer time that AI saves is profit for Eudia itself.

So the company has a very concrete reason to make its tools fit the way lawyers actually work, and the people who understand that work best are not engineers.

## Two roles, two different questions

The table below compares the two roles based on Eudia's job postings. The difference lies in the question each person must answer, not just in the title.

| Aspect | Forward Deployed Engineer | Forward Deployed Lawyer |
|---|---|---|
| Section on the careers page | Engineering | Product |
| How Eudia describes the role | A bridge between client stakeholders and the engineering team | The "eyes and ears" of the product and engineering teams in the field |
| Clients and location | Legal teams at F500 companies, defence organisations; the government version is in Washington, DC or remote | The government version works on site at DoW clients |
| Hiring requirements | Not specified in the quoted description | At least 4 years of legal practice; the government version prefers candidates with an existing security clearance |
| What they bring to the product | Connecting stakeholder needs with the engineering team | Deep understanding of how legal work actually happens |
| Core question (paraphrased) | What can the system do for this client? | How does this work really happen in practice? |

That the FDL sits in Product rather than Legal shows fairly clearly what Eudia wants from this lawyer: product requirements, not legal advice for the company. The government version also requires the FDL to stay in close contact with the growth, product and engineering teams, so that what they see in the field feeds back into the platform.

**Key point:** In regulated industries, a deployed engineer needs a domain expert sitting right beside them.

## One set of contracts, two ways of scoring

Consider a hypothetical scenario. You are an FDE deploying a contract-review agent for the legal team at an F500 company. The test set has 200 contracts, and the agent extracts 95% of the clauses correctly. On the dashboard, that figure looks very good.

A lawyer sitting next to you will ask a different question: where is the 5% that is wrong? Suppose the errors cluster in limitation-of-liability or change-of-control clauses, the kind of clause that can decide an M&A deal.

Then the reviewer has to reread everything, and the 95% saves almost no hours at all. If the errors are only in contact details, however, the client may accept them straight away.

This is exactly where each side has what the other lacks. An engineer cannot rank the severity of each type of legal error alone. A lawyer cannot build an eval set, a pipeline or a logging system alone. The best eval comes from the lawyer assigning a risk level to each error type and the engineer turning those levels into metrics that run on every deploy.

So the first thing to do at a client is to spend a session with the domain expert, before optimising any metric. Sort the errors with them into two groups: "acceptable" and "breaks the whole workflow". That list is worth more than any aggregate accuracy figure.

## When credentials and access outweigh skills

The government version adds one more criterion: candidates who already hold a security clearance are preferred. It is not a hard requirement, but together with the four years of practice, it shows that in this market credentials and access count for far more than an ordinary plus.

Someone who has never handled matters independently, or who lacks clearance, would probably struggle to be trusted to work on site at a defence client.

You do not need to become a lawyer. But you do need to understand why the person working beside you has the right to say no. When a lawyer says a result is "unusable", that is usually a professional risk assessment, not a personal preference.

A good FDE in this model turns that judgement into test cases rather than arguments. Every time the expert rejects a result, you add a case to the eval set, along with the reason they gave. After a few weeks, the eval set holds their domain experience, and nobody has to restate the criteria in every meeting.

## What does this mean for your career?

Eudia's logic is not limited to law. Any industry with outputs that require a licensed person to take responsibility may need engineers and domain experts working side by side in the same way. For anyone looking for an FDE job, this is a criterion for judging a job description.

When reading a job posting, check whether the company has a domain-expert role working alongside FDEs, and where the two roles report. In interviews, ask directly: who decides whether a result passes? The answer tells you whether you will spend your time building systems or arguing about what passing means.

On a CV, the line "built contract extraction pipeline" says little. State which expert you worked with to define the pass criteria, and which metric reflected the client's real risk. If you have no such project yet, practise on your own: take a public set of sample contracts, write your own scoring criteria, then ask someone in the field to correct them.

Eudia puts engineers in Engineering and lawyers in Product, but wants both to bring the field back to the product. In industries where one wrong clause can change an entire deal, the FDE who knows how to work with the person who decides what is "correct" is the one clients will want to keep.

**Try this week:**

- Take an eval set you are using and ask someone from the client's industry to sort the errors into two groups: 'acceptable' and 'breaks the whole workflow'. Then recompute your metrics by those two groups.
- Rewrite one line of your CV: state which expert you worked with to define the pass criteria, and which metric reflected real risk.
- Read three FDE job descriptions in tightly regulated industries. Note which companies have a domain-expert role working alongside FDEs and which department that role reports to.

## Sources

- [Forward Deployed Lawyer – Eudia (Greenhouse)](https://job-boards.greenhouse.io/eudia/jobs/4234525009)

- [Forward Deployed Lawyer - Eudia for Government (Greenhouse)](https://job-boards.greenhouse.io/eudia/jobs/4301304009)

- [Forward Deployed Lawyer - Eudia for Government (General Catalyst job board)](https://jobs.generalcatalyst.com/companies/eudia-2/jobs/84594749-forward-deployed-lawyer-eudia-for-government)

- [Forward Deployed Engineer – Eudia (General Catalyst job board)](https://jobs.generalcatalyst.com/companies/eudia-2/jobs/76581027-forward-deployed-engineer)

- [Eudia careers board (Greenhouse)](https://job-boards.greenhouse.io/eudia/jobs/4234585009)

- [Eudia Bags Up To $105m For 'Augmented Intelligence' Platform](https://www.artificiallawyer.com/2025/02/13/eudia-bags-up-to-105m-for-augmented-intelligence-platform/)

- [Eudia Opens 'AI-Augmented Law Firm' For M&A](https://www.artificiallawyer.com/2025/09/03/eudia-opens-ai-augmented-law-firm-for-ma/)

- [Eudia (company) – Wikipedia](https://en.wikipedia.org/wiki/Eudia_(company))