The forward deployed engineer: why AI's hottest new role is so hard to hire, and how to do it well

Juan Molina ยท

Postings for engineers who build AI inside customers' systems have grown several-fold. What the role is, what to look for, and how to interview for it.

Companies can build an impressive AI demo in a week. Making it work inside a real customer's systems, data and workflows is a different job, and it has created one of the fastest-growing roles in tech: the forward deployed engineer (FDE).

The growth figures vary by source, but all point the same way:

  • Indeed data reported by Business Insider: 5,330 FDE postings in April 2026, up from 643 a year earlier, a 729% increase.
  • Bloomberry's analysis of about 1,000 postings: a 1,165% rise between 2024 and 2025.

Both are cited in Scaler's analysis, updated October 5, 2026. This month, Digitimes reported that global tech companies are widening AI hiring "beyond model researchers" to frontline engineers who work inside customer operations (Digitimes, October 12, 2026).

The absolute numbers are still small. A September census by Plank counted 1,333 live roles with the FDE title at 565 companies, rising to an estimated 18,000 to 25,000 if similar customer-embedded engineering roles with other titles are included (Scaler). Small pool, fast-growing demand: that's what makes the role hard to hire.

What a forward deployed engineer actually does

An FDE writes production code inside a customer's environment rather than only for their own company's product. The role combines real engineering depth with constant customer contact. According to Bloomberry's analysis of postings:

  • 68% of roles involve on-site travel, and none carry a sales quota.
  • The most requested skills are Python (66%), TypeScript (35%), AI agents (35%), LLM experience (31%) and cloud platforms.
  • Successful candidates come from varied backgrounds: software engineering (45%), solutions engineering and architecture (22%), and data engineering and data science (15%).

What to look for

The FDE is part engineer, part consultant. That makes standard engineering filters a poor fit. Look for:

  • Building under real constraints. Experience integrating AI into messy, existing systems, not only greenfield projects.
  • Understanding before changing. Candidates who learn a customer's workflow before proposing a solution.
  • Comfort with ambiguity. The ability to break down a problem when there's no clear specification.
  • Translating for executives. Explaining technical decisions in terms of business outcomes.
  • Talking about outcomes. Strong FDE candidates usually describe past work by what changed for the customer, not only by the code they wrote.

How to interview for it

Algorithm puzzles tell you little about this role. A more useful loop tests the work itself:

  1. A practical integration task. Clean messy data or wire up a real-world integration.
  2. System design under constraints. For example, a legacy database or a strict security model.
  3. An underspecified case. Give a vague customer problem and watch how the candidate asks questions and breaks it down.
  4. A stakeholder conversation. Role-play a disagreement with a customer or an executive.

Since many strong FDEs come from solutions engineering or data roles rather than pure software engineering, widen your search beyond the job title. Some of the best candidates have never held the FDE title at all.

The short version

The forward deployed engineer exists because AI only creates value when it works inside real organizations. The role is growing fast, the talent pool is small, and standard hiring methods miss the qualities that matter most. The companies that hire them well:

  • define the role clearly;
  • look beyond the title;
  • interview for the actual work: building, listening and explaining.

If you're planning to hire forward deployed engineers or other customer-facing AI roles and want to talk it through, you can book a 15-minute call.

Not ready for a call yet? You can post your open role for free.