How to Hire Your First FDE

Your first forward deployed engineer is one of the most leveraged hires you will make. They will sit in front of your most important customers, write code you will not review line by line, and tell your product team what to build next. Get it right and your deployments speed up while your product gets sharper. Get it wrong and you have an expensive engineer who is either hiding from customers or charming them while nothing ships.

This guide covers what we have learned from doing the job and from hiring for it: how senior to go, where to find people, how to run an interview loop that actually tests for the role, and the mistakes we see most often.

Before You Post Anything: Define the Job

Start by writing down what this person will do in their first 90 days. Not the job description. The work.

• Which customers will they own?
• What does “live” mean for those customers?
• How much time on site, and where?
• Who do they report to, and who acts on their product feedback?
• Will they do any pre-sale work, and if so, how much?

This sounds obvious. It is the step most teams skip. LeadDev’s piece on the rise of the FDE quotes a former Palantir FDE asking whether companies can even define the role, and a Palantir executive describing many copycat FDE programs as a half measure. Henley Wing Chiu’s analysis of about 1,000 FDE job postings found that a large share of postings titled FDE were really sales engineering or internal tools roles. If you are not sure which role you need, read FDE vs Solutions Engineer vs Sales Engineer first.

How Senior Should Your First FDE Be?

Hire senior for your first one. Not necessarily a staff engineer, but someone who has shipped real systems and has been in front of customers before.

Here is why. Your first FDE will not have a playbook, a manager who has done the job, or a peer to ask. They will make judgment calls daily about what to build, what to push back on, and when to escalate. Posting your first FDE role at junior or mid level is one of the most common mistakes we see, for exactly this reason.

Look at what the companies that do this at scale ask for. OpenAI’s FDE postings ask for five or more years of engineering or technical deployment experience, including customer-facing work. Anthropic’s ask for four or more years in a technical, customer-facing role, and explicitly welcome former technical founders. The Pragmatic Engineer notes the range is wide: Ramp prefers five or more years for senior FDEs, while Palantir hires people about a year out of college. Palantir can do that because it has a deep bench of experienced FDEs to train them. You do not, yet.

Stage matters too. At seed, you want a scrappy generalist who can own an integration end to end. At Series A, look for someone with a few deployments of pattern recognition. Later, a domain specialist starts to pay off, especially in regulated industries.

What to Look For

Three things, and you need all three:

1. Production engineering. They can ship working code in a codebase and environment they have never seen. Not just prototypes.
2. Customer judgment. They can run a discovery conversation, read a room, say no to a customer without losing them, and explain technical trade-offs to an executive.
3. Ownership of outcomes. They talk about past work in terms of what changed for the customer, not which framework they used.

a16z partner Joe Schmidt, in Trading Margin for Moat, says to hire “curious hustlers” with high agency and a healthy disregard for the status quo, and notes that a PhD is not required. Industry exposure helps. If your customers are hospitals, an engineer who has worked with healthcare data will be faster from day one.

Where to Find Them

Good FDEs are rarely on job boards. They are busy at customer sites. Expect to source passively. Posting the job and waiting for applicants rarely works for this role.

The best pools we know:

• Former Palantir FDEs. The original training ground. Also the most heavily recruited.
• FDEs and applied AI engineers at AI labs and AI startups. They have done this recently, on LLM products.
• Implementation, integration and solutions engineers at enterprise software companies. Titles are unreliable here. Many people doing true FDE work are called something else.
• Former technical founders. They have done discovery, built under ambiguity and sold to skeptical buyers.
• Data engineers with client-facing experience. LeadDev quotes a recruiter who has seen data engineers make the move successfully.
• Referrals from FDEs. In our experience, the highest-yield channel. Good FDEs know other good FDEs.

Search for the skill set, not the title. Someone titled “Implementation Engineer” who writes Python against customer APIs every day is closer to an FDE than a “Forward Deployed Engineer” who runs demos.

The Interview Loop

The most common mistake is running your standard software engineering loop. Algorithm puzzles do not tell you whether someone can untangle a customer’s broken data pipeline, and a normal loop never tests the customer half of the job at all.

Design the loop around the actual work. Here is a sample loop with timings. Adjust it to your product.

A Sample Loop

1. Hiring manager screen (30 minutes). Walk through their best deployment. Who was the customer? What did they promise? What broke? What did they build? What would they do differently? You are listening for outcomes and specifics. A useful question from Natalie Meurer of Sierra, in her talk The Dirty Secret of Forward Deployed Engineering: ask what vintage of FDE they are, meaning which kind of forward deployed work shaped them.

2. Practical build (90 minutes live, or a take-home capped at 3 to 4 hours). Give them messy, realistic data and a half-documented API, and ask them to wire up something that works. For an AI product, include an evaluation step. Score working output, judgment about what to skip, and the note they write explaining their choices. Skip algorithm puzzles. They tell you little about this job.

3. Decomposition or scoping case (60 minutes). Hand them a vague customer problem. “Our claims team is drowning and we want AI to help.” Let them ask questions while an interviewer plays the customer. Halfway through, reveal a constraint, such as no access to production data for six weeks. You want to see them break the problem down, find the real need, and re-plan without panic. Palantir lists technical decomposition as a core FDE skill in its Dev versus Delta post.

4. Customer role-play (45 minutes). An interviewer plays a frustrated customer executive whose go-live slipped. The candidate has to de-escalate, find out what is really wrong, and agree on a next step without overpromising. Our role-play script gives you a ready-made scenario.

5. Team and values conversation (45 minutes). Meet the account executive and a product engineer they would work with. Test how they hand off feedback to product and how they handle a disagreement with sales over scope.

That is roughly four and a half hours of candidate time, which is fair for a role this important. Score every stage against the same criteria. Our interview scorecard breaks this down, and our interview questions for FDEs gives you more prompts.

Red Flags

• They say they would rather just focus on the code, and they see customers as a distraction.
• They are great with customers but cannot ship working code. That is diagnosis without a fix.
• Every past project ended at the demo or the POC.
• They agree with everything the role-play customer asks for.
• They cannot name a time they told a customer no.

The Mistakes We See Most

Hiring a consultant who cannot code. Strong communicators who cannot build will win the first meeting and stall the deployment.

Hiring a great engineer who avoids customers. The reverse mistake. They will build the wrong thing beautifully.

Using your FDE as a demo engineer. Pre-sale scoping is fine in small doses. A standing demo duty pulls them away from the customers who already pay you.

Giving them a quota. a16z recommends keeping FDEs closely aligned with account executives but without quotas. Their incentive should be the customer’s success.

No path for their feedback. If what your FDE learns in the field never reaches the roadmap, you are paying for half the role. a16z’s essay stresses tight feedback loops from FDEs straight to the product team.

Overloading them. One person cannot run deep deployments at ten customers. See how many FDEs per customer for how to think about load.

No onboarding plan. Even senior FDEs need to learn your product cold before they represent it. Our guide to a new FDE’s first 90 days lays out a plan.

Where We Come In

We only place FDEs, and the business is run by a former FDE. Every candidate we introduce has already passed a technical screen and a live customer role-play, so your loop can focus on fit with your product and your customers. You run the final interviews, and interviewing our candidates is free. Read how we vet or start hiring when you are ready.