Skip to content
Search

Forward Deployed Engineer

lastmod 2026-03-04created 2026-03-022 edits

A forward deployed engineer (FDE) is an engineer deployed directly on-site with customers to implement the ‘last mile’ of a technically complex product.

The complete guide to hiring FDEs

About

  • A forward deployed engineer (FDE) is an engineer deployed directly on-site with customers to implement the ‘last mile’ of a technically complex product
  • Not just a consultant, but a developer who writes and debugs real production code and uncovers new product opportunities
  • This model does not suit every company; it works only when three conditions are met: large-contract customers, a non-uniform ICP, and openness about product direction
  • The best FDEs are early-career people with independent thinking, grit, strong technical skills, and business curiosity, the type who does not rely on existing playbooks
  • The key to running an FDE team is focusing on the hardest problems of the biggest customers while combining on-site deployment with proper scope management

Origins and present of the FDE

  • Palantir co-founder Alex Karp first devised the Forward Deployed Engineer (FDE) role about 20 years ago, inspired by the observation that waiters in French restaurants are an extension of the kitchen staff
  • An FDE is deployed directly on-site with customers to build the "last mile" of a product in production; unlike traditional solution consultants or sales engineers, they write and debug real production code
  • For a long time there was skepticism that it was "just high-end consulting," but interest in the role reignited after Palantir's market cap passed $300 billion
  • FDE job postings grew 800% between January and September 2025, and AI startups including OpenAI are building FDE teams to strengthen enterprise sales

Where FDEs create value

  • An FDE is not just an implementation role but a real member of the software engineering team, who must build software directly while talking with customers every day
    • Serval's FDEs actually shipped more than 60 third-party app integrations, an agent performance feedback system, an SLA system, and more as product
  • The scope of Palantir FDE work ranges from reducing manufacturing defect rates to deploying software for managing natural disaster relief supplies
  • Helping close large deals
    • Looker used FDEs in sales by offering prospects a free trial on their real data along with strong up-front implementation support
      • Co-founder Lloyd Tabb: "Not deciding between product and service opened a third way. We sold it as a product, but forward deployed during the free trial so it felt like a custom service"
      • Always built the PoC on the prospect's real dataset instead of a dummy demo version
    • At Palantir too, there was a case where 3–4 FDEs closed a deal for an energy-sector customer by directly building a solution tailored to the problem, without a roadmap or headquarters approval
  • Finding product opportunities by embedding with customers
    • On-site embedding can surface deep insights that are hard to get from a 30-minute Zoom call
      • "Living on-site with the customer is the core of the FDE. It's not setting up user interviews, but prototyping what you heard that day and showing it the next day"
    • Some of the most successful FDE cases came from areas unrelated to Palantir's core product
    • However, there is a risk that FDEs randomly build features unrelated to improving the core product, so product sense is needed to identify problems that can serve other customers as well
  • Scaling early founder energy
    • The FDE model is a way to recreate and scale the early co-founder energy of a CTO who heard customer feedback directly and fixed it in code right away
      • At Serval, the difference between FDEs and regular engineers is small; FDEs spend about 20% of their time with customers and focus on building product capabilities rather than infrastructure
    • In the traditional feedback-to-product cycle, going from solutions engineer → PM → engineering manager → sprint scheduling takes months, but the FDE model removes these intermediate steps
  • Clearing low-priority features
    • FDEs can actually handle P2-level feature requests that keep getting pushed down the roadmap
      • At Verkada, the sales team piled customer requests into a Slack channel called "Feature Garage", but most were left untouched
    • "If P2s keep piling up, you end up with an inferior product even if you only focused on what was technically right. In the FDE model, P2s actually get looked at"

Self-assessment before adopting FDEs

  • Investing in an FDE team is costly, especially for early-stage startups, and can burn through money quickly if the business model math doesn't work
  • Looker CEO Frank Bien decided to invest in FDEs after validating a model in which $100 million ARR was achievable with 2,000 customers
    • "If you fully understand your model, you can check by calculation whether it can support all the resources you put in. But if we had been unsure whether reaching $100 million needed 2,000 or 100,000 customers, we would have been burning VC money"
  • Condition 1: You have secured or are targeting large customers
    • FDE is inherently an upmarket strategy; if a PLG freemium model is your end state, FDEs are not a good fit
    • Ironclad understood from the start that the legal departments of huge global companies were its ideal customers, and adopted FDEs on the judgment that custom implementation was needed to win enterprise ACV
  • Condition 2: You do not have strong opinions about how the product should be used
    • If the company has strong opinions about what the product should look like in the future, a PM or customer-facing engineer is a better fit than an FDE
    • At one end of the spectrum is Apple (every user gets the same experience); at the other is Palantir (a platform that can be reshaped to fit diverse problems and organizations)
      • In Palantir's early days, people rarely said "the product should be like this"; FDEs learned from concrete use cases and customers and gradually built a valuable product
  • Condition 3: Your ICP is not uniform
    • If customer criteria are defined in great detail, the need for an FDE team may shrink
    • Ironclad's first 50 customers had almost nothing in common: public tech companies, YC startups, global beauty brands, pro sports teams, and so on
      • A Japanese-language influencer contract workflow and an MLB season ticket sales contract had completely different needs
    • Promise serves the US government as a customer and is building an FDE team because of a heterogeneous customer environment where each state runs its programs differently
  • Don't force the FDE into an engineering or post-sales role
    • Many founders say they want FDEs, but often what they actually need is a regular software engineer, an implementation consultant, or a CSM
    • Validation questions recommended by Tiffany Siu of First Round:
      • "What prompted you to open this position? Who is doing this work now? What happens if you don't hire this person?"
      • "What exactly does this person's day-to-day work look like?"
      • "What are this person's success metrics?"
    • If you "want the FDE to close X deals or run X demos," you need a role closer to sales, not an FDE

How to hire the right FDE

  • No need to fixate on finding people who already hold the FDE title. FDE role structures differ from company to company, so check the actual work rather than the title
  • Five traits the best FDEs share
    • People who don't rely on playbooks (early-career people have an edge): At Palantir, new graduates made up a large share of FDEs. Independent thinkers who believe they can solve any problem, rather than pattern-matching, are a good fit
      • Candidates with overly fixed ways, like 10+ years of FAANG experience, are actually a red flag
    • Grit: FDE work deals with extremely hard problem spaces, so a "willingness to endure pain" is essential. Ironclad's best FDEs were also described as "grinders"
      • FDEs act as mini founders within the organization and need strong product sense and a genuine interest in sales
    • High technical bar: Ironclad's best FDEs are at the level of staff engineers at top tech companies. At Palantir, too, FDEs had to pass the same interview process as software engineers
    • A habit of constantly building: The best FDEs are "compulsive builders," the type who can't help making things, whether building tools, shipping apps, or contributing to open source
    • Deep curiosity about business: People who get energy from how businesses work, such as digging deep into the legal risks of influencer marketing
  • Interview design
    • Instead of typical big-tech coding interviews, Palantir runs high-level, problem-solving interviews centered on real customer problems
      • E.g., after explaining insider trading, asking "What data would you need, what questions would you ask the customer, and what would you look for?" to evaluate both business reasoning and technical reasoning
    • Ironclad uses an open-ended presentation interview in which candidates present a problem from their own career and teach how they solved it with technology
      • A representative example showed photos of a physical 'deal room' for aviation finance transactions with thousands of pages of sticky-note work, followed by the Excel macro that later automated it

Scoping the FDE role: 3 tactics for making your first FDE team impactful

  • Deploy FDEs on the hardest problems of your biggest customers
    • Ironclad sent FDEs to every customer early on, but as it grew it switched to assigning FDEs only to high-ACV customers
      • Rather than "either a full FDE company or not at all," it systematized a menu of options according to customer needs and standardized implementation for the down-market
    • Serval also deployed to every customer at first, but now prioritizes large customers, such as companies with 1,000+ employees
  • The importance of on-site deployment
    • FDEs do their best work on-site. Going beyond the list of requirements the customer specified in the contract, they share the day-to-day experience and can make discoveries that were not in the contract
      • The on-site visit culture of Ironclad FDEs was so well known that a Fortune 100 general counsel called them "The Backpacks"
  • A balanced approach to scope creep: accept it, but beware of ‘covering product problems with human labor’
    • Embrace the service nature of the FDE role, but avoid pouring unlimited time into solutions that won't benefit future customers
    • Early Ironclad treated scope creep as a signal that "there are more problems the customer can solve," and thanks to workflow-based pricing, the software economics benefited from scope expansion
    • Bad scope creep is endless repetitive work on a workflow with a limited number of users; if no software upside is visible, the work should stop
    • A true FDE is so absorbed in solving problems that they don't even notice going beyond the contract scope, so scope management is the manager's job

See also

author Zer0 sitemap