AI Forward Deployed Engineer

Every few months a new job title gets buzzy in Silicon Valley. The latest one I keep getting asked about is the AI Forward Deployed Engineer, or FDE. The AI researcher Andrew Ng recently wrote about it, and what he said is good news for anyone worried that AI is only here to take jobs away. So let me explain what an FDE actually does, why the role is surging, and how it fits into the much larger wave of AI jobs being created right now.

What is an AI Forward Deployed Engineer?

An FDE is an engineer who is embedded inside a client company to help build and customize AI solutions for that specific business. Rather than building a product for everyone, an FDE shows up at one company and tailors the technology to that company’s real needs. A lot of that work today means taking an off-the-shelf large language model and turning it into a custom agentic workflow (an AI system that can take a series of actions on its own to finish a job), tuned for how that business actually operates.

Where did the FDE role come from?

It is not brand new. The data company Palantir pioneered the role about two decades ago, sending its engineers out to government sites to work directly on secure, air-gapped networks (closed systems with no outside internet connection). The idea was simple: instead of guessing what the client needed from afar, put a skilled engineer in the room with them. That same idea is now being applied to AI.

Why is the FDE role surging now?

Two reasons. First, Ng notes that big AI labs, including Anthropic and OpenAI, have started building teams to place FDEs inside client organizations. When the makers of the models start hiring for a role, the rest of the industry pays attention. Second, there is simply a lot of work involved in turning a general-purpose model into something that fits one company’s processes. That gap, between a raw model and a working business tool, is exactly what an FDE is hired to close.

What skills does an FDE need?

This is the part beginners often miss: an FDE is not only a coder. Because they sit with clients, the role rewards a mix of skills:

• Technical skill to build and tune AI systems and agentic workflows.
• Communication skill to understand what a client really needs and explain complex technology in plain terms.
• Business sense to prioritize the projects that matter most.
• The confidence to push back respectfully when a client asks for something that is not realistic.

If you enjoy both building things and talking to people, that combination is rarer than you would think, and valuable.

FDE versus AI Engineer: which has more jobs?

Here is Ng’s key point, and where the real opportunity is. He believes there will be far more AI Engineer jobs than FDE jobs. A company might accept a handful of embedded FDEs, but it will usually want many more of its own employees building its AI projects. Ng even notes that his own organizations hire some FDEs but hire far more AI Engineers.

There is also a trust issue. An FDE is usually there to integrate one vendor’s product deeply into a company. That makes some clients nervous, because tying your processes tightly to a single vendor reduces your optionality, your freedom to switch to whatever AI service turns out to be best a year from now. In a fast-moving field, that flexibility is worth a lot.

	AI Forward Deployed Engineer	AI Engineer
Where they work	Embedded inside a client company	On their own company’s team
Main job	Customize a vendor’s AI for one client	Build AI features into the company’s own products
Number of jobs	Fewer (a few per client)	Far more (Ng’s view)
Vendor tie	Tied to one vendor’s product	Usually vendor-neutral

What does an AI Engineer actually do?

Ng describes surging demand for AI Engineers who can build software using AI components, things like prompting large language models, wiring up agentic frameworks, and writing evals (short for evaluations, the tests that measure whether an AI system is actually doing its job well). Just as important, good AI Engineers know how to work alongside AI coding agents like Claude Code, Codex, and OpenCode, which write and edit code on your direction. The skill is less about memorizing syntax and more about directing these tools well.

What new AI jobs might come next?

Ng expects the AI Engineer role to split into specialties over time, the same way the old generic “software engineer” job split into frontend, backend, mobile, data, and devops roles. Nobody knows the exact names yet, but he floats a few possibilities:

• AI FDEs, the embedded client role described above.
• LLMOps Engineers, who keep AI models running reliably in production.
• Evals Engineers, who specialize in testing AI quality.
• AI Data Engineers, who prepare the data that AI systems learn from and use.
• Harness Engineers, who build the scaffolding that lets AI agents work safely.

Plenty of roles here do not even have settled names yet, which is part of what makes this an exciting time to start.

Does this mean AI is creating jobs, not destroying them?

That is Ng’s larger argument, and it is one we share at Beginners in AI. The fear that AI will cause a sudden job-market collapse, sometimes called the “jobpocalypse,” is not playing out. New roles like the FDE are appearing precisely because AI needs skilled people to shape it, point it at the right problems, and check its work. AI is a powerful tool, but it still takes human judgment, communication, and business sense to make it useful. The opportunity is not to compete with AI, but to become the person who knows how to wield it.

Skilled AI Engineers are in very high demand!

Andrew Ng, The Batch

How do you start down this path?

You do not need a Silicon Valley job to begin building these skills. Learn to prompt a model well, try wiring a simple agentic workflow, and practice directing an AI coding agent on a small project. Work on your ability to explain technical ideas simply, since that is what separates an FDE from a pure coder. Our AI for professions guide and beginner path are good places to start, and the best AI assistants roundup helps you pick a tool to practice with.

Common questions

What does FDE stand for?

Forward Deployed Engineer. It is an engineer placed inside a client company to build and customize AI solutions for that specific business.

Is an FDE the same as an AI Engineer?

No. An FDE works embedded at a client and integrates a vendor’s AI; an AI Engineer usually builds AI into their own company’s products. Ng expects far more AI Engineer jobs.

Do you need to be a strong coder to be an FDE?

Coding matters, but so do communication and business skills. FDEs talk to clients, prioritize work, and explain (and sometimes push back on) technical decisions.

Will AI replace these jobs?

These roles exist because AI needs skilled people to shape and check it. Ng argues AI is creating new jobs, not causing a job-market collapse.

What should a beginner learn first?

Prompting models, basic agentic workflows, evals, and how to work with AI coding agents like Claude Code. Clear communication is a real advantage.

Sources

• Andrew Ng, The Batch (DeepLearning.AI)
• Anthropic: Claude Code
• DeepLearning.AI (AI courses)

Last reviewed: June 2026. Based on Andrew Ng’s writing in The Batch; job-market views are his analysis.

You may also like

• AI for Professions: how AI changes your job
• How to use Claude AI
• How to install Claude Code
• The AI Tools Directory
• Start Here: the beginner path
• The AI Glossary
• The daily Beginners in AI newsletter