How do a Forward Deployed Engineer and an AI Engineer differ?

Authors
How do a Forward Deployed Engineer and an AI Engineer differ?

A Forward Deployed Engineer builds and ships AI software for one customer at a time inside that customer's environment, while an AI Engineer builds one AI product for many users inside their own company's codebase.

In this blog, we will learn about how a Forward Deployed Engineer and an AI Engineer differ. We will also see what each role owns, where each one sits, what a typical week looks like for each of them, what each one builds with a coding agent, how the two roles compare side by side, and which one to pick based on our situation.

We will cover the following:

  • Who is an AI Engineer?
  • Who is a Forward Deployed Engineer?
  • Forward Deployed Engineer vs AI Engineer
  • Which one should you pick?

I am Amit Shekhar, Founder @ Outcome School, I have taught and mentored many developers, and their efforts landed them high-paying tech jobs, helped many tech companies in solving their unique problems, and created many open-source libraries being used by top companies. I am passionate about sharing knowledge through open-source, blogs, and videos.

I teach AI and Machine Learning at Outcome School.

Let's get started.

Who is an AI Engineer?

An AI Engineer is an engineer who builds products on top of existing AI models. The models are built by someone else. The AI Engineer uses them through an API and builds a product around them.

So, an AI Engineer does not train a model from scratch. The AI Engineer takes an LLM (Large Language Model) and connects it to the data, the tools, and the users of a product.

What does an AI Engineer own?

An AI Engineer owns one product, or one feature inside a product, that is used by many users.

Let's say a company has a customer support product. The AI Engineer owns the AI assistant inside that product. The assistant must answer questions from thousands of customers, and it must do that every day without breaking.

So, the AI Engineer owns the following:

  • The prompts that tell the LLM how to behave.
  • The RAG (Retrieval Augmented Generation) pipeline that fetches the right documents before the LLM answers.
  • The agent that uses tools, like searching an order or creating a ticket.
  • The evaluation that measures how good the answers are.
  • The deployment that keeps it running for every user.

Where does an AI Engineer sit?

The AI Engineer sits inside their own company, inside their own codebase, and inside their own environment. The engineer knows the code, the data, and the infrastructure, because the team built all of it.

The AI Engineer rarely talks to a user directly. The product manager brings the requirements, and the AI Engineer builds for everyone at once.

What does a typical week look like?

A typical week of an AI Engineer looks like below:

  • Reading the evaluation results from last week and finding the cases where the assistant failed.
  • Changing the prompt or the retrieval step to fix those cases.
  • Directing a coding agent to build a new tool for the agent, then reviewing and testing what it produced.
  • Running the evaluation again to confirm the fix did not break something else.
  • Shipping the change to all users and watching the cost and the response time.

The same product gets better week after week.

A quick note for you

No matter which tech domain you work in, get familiar with these topics:

  • LLM
  • RAG
  • MCP
  • Agent
  • Fine-tuning
  • Quantization

We put it all together in one video:

AI Engineering Explained: LLM, RAG, MCP, Agent, Fine-Tuning, and Quantization

No need to stop reading - bookmark it and watch later when you get time. Future you will thank you.

Now, let's get back to the topic.

What does an AI Engineer build with a coding agent?

The AI Engineer builds with a coding agent, such as Claude Code. The engineer describes the task to the coding agent, and then reviews, tests, and ships the result.

For example, the AI Engineer asks the coding agent to write a retrieval step that splits the documents into chunks, stores them in a search index, and returns the top matches for a question. The coding agent writes it. The AI Engineer checks the chunk sizes, runs the evaluation, and decides whether the result is good enough for all the users.

This was about the AI Engineer. Now, it's time to learn about the Forward Deployed Engineer.

Who is a Forward Deployed Engineer?

A Forward Deployed Engineer is an engineer who is embedded with one customer and who scopes, builds, and ships production AI software inside that customer's environment.

Let's understand the term Forward Deployed. In the army, forward deployed means stationed close to the front line, not at the headquarters. The idea is the same here. The engineer is not sitting at their own company. The engineer is sent forward, to the customer, and works from there.

Palantir created this role. Now, it is common at AI companies, because an AI product rarely fits a customer without extra work. Someone has to go in, connect it to the customer's data and systems, and make it work.

Want to become a Forward Deployed Engineer?

These are the skills this role needs, and we teach them in our AI and Machine Learning Program at Outcome School:

  • RAG
  • AI Agent
  • Model Context Protocol (MCP)
  • Evaluation of LLMs and Agents
  • Model Deployment and Serving
  • AI Security

We take you from the fundamentals to building real AI systems, step by step.

What does a Forward Deployed Engineer own?

A Forward Deployed Engineer owns the outcome for one customer at a time.

Let's say an AI company sells an agent platform. A bank buys it. The bank has its own databases, its own documents, its own rules, and its own security team. The platform alone does nothing for the bank until it is connected to all of that.

So, the Forward Deployed Engineer goes to the bank and owns the following:

  • Scoping: understanding what the bank needs, and saying no to what is not possible.
  • Integrations: connecting the platform to the bank's data and systems.
  • The agent itself: building the prompts, the retrieval, and the tools for the bank's use case.
  • Delivery: shipping it into the bank's production and staying until it works.
  • Communication: explaining progress, trade-offs, and problems to the people at the bank.

Scoping and communication are half of the job.

Where does a Forward Deployed Engineer sit?

The Forward Deployed Engineer sits inside the customer's environment. The codebase is not theirs. The data is not theirs. The infrastructure is not theirs. The engineer must learn all of it quickly and must work within the customer's rules.

The Forward Deployed Engineer talks to the customer every day, often in the same room or on the same call.

Note: This role is different from a Solutions Engineer. A Solutions Engineer supports sales and builds demos before the deal. A Forward Deployed Engineer owns the delivery after the deal, in production.

What does a typical week look like?

A typical week of a Forward Deployed Engineer looks like below:

  • Meeting the customer to agree on what will be delivered this week.
  • Studying the customer's systems and data to find how to connect to them.
  • Directing a coding agent to build the connector and the agent itself, then reviewing and testing it inside the customer's environment.
  • Showing a working version to the customer and collecting feedback.
  • Fixing the issues the customer found, and writing down what was learned for the product team at the AI company.

When one customer is done, the next customer begins, and the week starts again from scoping.

What does a Forward Deployed Engineer build with a coding agent?

The Forward Deployed Engineer also builds with a coding agent, such as Claude Code. The difference is in what is built and where.

For example, the Forward Deployed Engineer asks the coding agent to write a connector that reads records from the bank's internal database and converts them into the format the agent platform expects. The coding agent writes it. The Forward Deployed Engineer runs it against the bank's real data, checks the records that failed, and fixes the rare cases that only this bank has.

Here, the coding agent speeds up the coding. The scoping and the verification inside an unfamiliar environment are still the engineer's job.

Stay updated: Subscribe to our newsletter to get our latest AI and Machine Learning blogs straight to your inbox.

Forward Deployed Engineer vs AI Engineer

Both the roles use the same core skills: LLM APIs, RAG, agents, evaluation, and agentic coding. The difference is in the ownership and the setting.

Let me tabulate the differences between a Forward Deployed Engineer and an AI Engineer for your better understanding so that you can decide which one to pick based on your situation.

AspectForward Deployed EngineerAI Engineer
Builds forOne customer at a timeMany users of one product
Works insideThe customer's environmentTheir own company's codebase
OwnsThe outcome for that customerThe product or feature for all users
ScopingDone by the engineer with the customerDone mostly by the product manager
Time with usersEvery day, directlyRare, mostly through data and feedback
CodebaseNew and unfamiliar for each customerKnown, built by the team
CommunicationHalf of the jobA smaller part of the job
Measure of successThe customer's system works in productionThe evaluation and the product metrics improve
Work rhythmProject by project, restarts with each customerContinuous improvement of the same product
TravelOftenRare
Core skillsLLM APIs, RAG, agents, evaluation, agentic codingLLM APIs, RAG, agents, evaluation, agentic coding

Which one should you pick?

We must choose based on our situation.

Pick the Forward Deployed Engineer role if:

  • We enjoy talking to people as much as building.
  • We like a new problem with every customer.
  • We are comfortable working in a codebase and an environment that are not ours.
  • We want to see the direct effect of our work on one customer.

Pick the AI Engineer role if:

  • We enjoy improving one product over a long time.
  • We prefer a known codebase and known infrastructure.
  • We like working from data and evaluation more than from meetings.
  • We want to build for a large number of users at once.

If we are a Software Engineer moving into AI, the AI Engineer role is the closer step, because the environment is familiar and only the AI layer is new. If we have worked in consulting or in a client-facing role, the Forward Deployed Engineer role uses that experience directly.

The skills we build for one role transfer to the other, and many engineers move between them.

Now we must have understood how a Forward Deployed Engineer and an AI Engineer differ.

Frequently Asked Questions

Does an AI Engineer train models from scratch?

No. An AI Engineer does not train a model from scratch. The models are built by someone else, and the AI Engineer uses them through an API. The AI Engineer takes an LLM and connects it to the data, the tools, and the users of a product.

Which role is the easier move for a Software Engineer going into AI?

The AI Engineer role is the closer step for a Software Engineer moving into AI. The environment is familiar, and only the AI layer is new. If we have worked in consulting or in a client-facing role instead, the Forward Deployed Engineer role uses that experience directly.

Can an engineer move between the Forward Deployed Engineer and AI Engineer roles?

Yes. The skills we build for one role transfer to the other, and many engineers move between them. Both roles use the same core skills, which are LLM APIs, RAG, agents, evaluation, and agentic coding. The difference is in the ownership and the setting, not in the skills.

Does a coding agent do the whole job of a Forward Deployed Engineer?

No. A coding agent such as Claude Code speeds up the coding, but the scoping and the verification inside an unfamiliar environment are still the engineer's job. The Forward Deployed Engineer runs the code against the customer's real data, checks the records that failed, and fixes the rare cases that only this customer has.

That's it for now.

Thanks

Amit Shekhar
Founder @ Outcome School

You can connect with me on:

Follow Outcome School on:

Read all of our high-quality blogs here.