dear [Forward Deployed Engineer]

What’s an FDE? The engineering role that grew 800% in a year.

Entry-level dev jobs are down. But forward deployed engineering postings are up 800%. It’s a real job, and it’s open to new grads.

A conversation from dear [Forward Deployed Engineer]Episode coming soon
Clara Shih with Leo Mehr, formerly Director of Engineering at Ramp.

What does an FDE actually do?

A forward deployed engineer (FDE) embeds directly with customers, writes production code, and owns the results. If you map out the problem on day one, you’re also the one fixing it when it breaks in production six months later.

It’s basically three jobs in one:

  • Consultant: Finding the highest-value problems in real workflows.
  • Product Manager: Scoping the work and feeding insights back to the core team.
  • Engineer: Shipping code inside someone else’s legacy systems.

Most FDE job descriptions ask for 3 to 5 years of experience, making it look off-limits for new grads. It isn’t. Companies like Palantir, who invented the model, and Ramp specifically hire recent grads. Why? Because they want people who can think on their feet, not just copy-paste from an old playbook.

The job is half coding and half talking to people. You can build both skills before you ever get officially hired.

Skills that stay valuable

  • Finding problems and identifying stakeholders
  • Scoping ruthlessly and shipping fast
  • Writing production code inside unfamiliar systems
  • LLM integration, retrieval, and model evaluation
  • Explaining technical tradeoffs to non-engineers
  • Catching errors in AI output
800%

Growth in FDE job postings in a single year

FDE Pulse, 2026
1,000

Forward deployed engineers Salesforce has committed to hiring

Salesforce, 2026
$130K+

Typical base-plus-bonus range, reaching $300K and higher at AI labs

FDE Pulse, 2026
Aaron Levie
“Forward deployed engineers, or equivalent, are about to become one of the most in-demand jobs in tech. And one of the most important functions for AI rollouts.”
Aaron Levie · CEO, Box

Half code, half conversation.

Leo Mehr

Leo Mehr built the forward deployed engineering team at Ramp, growing it from two people to 25 as Director of Engineering. His own early resume included reffing soccer games and working the floor at Office Depot. Dealing with everyday people in retail taught him how to navigate frustrated customers, deescalate problems, and explain things simply, the same skills that make for a strong FDE.

At Ramp, the job is about 80% software engineering and 20% product management, customer success, and sales. FDEs sit on the engineering team and have to pass the exact same technical interviews. The difference? FDEs actually focus on business problems and talking to customers.

Leo hired 10 new grads into the role. His ideal candidate profile is simple: Prove you can code, show you can hold a conversation, and demonstrate willingness to learn whichever side is weaker for you.

“If you have a non-technical background but you are excited to be more technical, FDE can be one of the best roles for you.”
Leo Mehr · Former Director of Engineering, Ramp

Checklist for entry-level FDEs.

It’s the classic catch-22: You need experience to get experience. But here is what counts as experience before anyone officially hires you.

  1. Ship something public that a stranger can actually use. A site, a tool, or a GitHub project with real users. Leo’s test is whether he can open it and form an opinion.

  2. A consistent commit history. A year of green squares on GitHub proves you code regularly, rather than just panic-coding right before a deadline.

  3. A people-facing job in retail, tutoring, TA work, or sales. Pair this with a technical project and you have the exact profile top FDE teams are looking for.

  4. Proof you can cut features to ship faster. Scoping is half the job. Explaining what you chose not to build shows better judgment than just showing off a finished project.

  5. Opinions on the AI tools you use. Companies like Ramp want you using AI during interviews because that’s exactly how the actual job works.

The best FDEs use AI at every step.

Discovery

Granola / Notion AI

Record customer calls and ask Claude to find the real problems. The value isn’t in taking the notes, it’s in what you extract from them.

Pattern finding

Claude

Scan a hundred conversations to see if anyone else raised the same issue. This used to be a product manager’s entire job.

Scoping

Claude / ChatGPT

Figure out if you should hack a quick fix for one customer or build a feature into the core product. Getting this wrong costs more than building slowly.

Building

Claude Code / Codex / Cursor

Turn a spec into a shipped feature in a single afternoon. Teams are doing in a day what used to take weeks.

Quality control

Your own judgment

AI can easily generate slop (18 pages when two would do). Brevity is a rare skill now, and no tool can do it for you.

Interviews

Whatever you actually use

Top companies want to see how you prompt and scope in real time. Companies that ban AI in interviews are going to get left behind.

Map your path to FDE

Get a free game plan that assesses what you already bring and the specific gaps to close.

Analyze my fit →