Aced (formerly Exponent) · Published 2026-07-02

What is a Forward Deployed Engineer? (with Founding Rippling FDE)

privateReview: unreviewedConfidence: high
forward deployed engineerenterprise softwarepalantirripplingconsultingforward-deployed-engineeringagentic-airequirements-elicitationiterative-deliveryenterprise-aicustomer-adoptionproduct-engineeringactive-listening
Open on YouTube ↗

Summary

Overview

  • Speaker: Stephen Cognetta and Kevin Bai
  • Channel: Aced (formerly Exponent)
  • Main topic: Forward Deployed Engineering (FDE)
  • Purpose: Provide a comprehensive guide on the Forward Deployed Engineer role, clarifying misconceptions, explaining the three core capabilities (consultant, product manager, software engineer), and offering guidance for candidates seeking FDE positions. Stephen Cognetta interviews Kevin Bai, the founding Forward Deployed Engineer at Rippling and former FDE at Palantir, to unpack the role of an FDE, how it differs from traditional software engineering and sales engineering, the key skills required, and the interview and recruiting process for FDE roles.

Topic Map

Introduction and Speaker Backgrounds

  • Explanation: Stephen Cognetta introduces Kevin Bai, who shares his background as the founding FDE at Rippling and former FDE at Palantir.
  • Key claims:
    • Kevin Bai was the first FDE at Rippling, joining over a year ago to build the function.
    • Kevin previously worked at Palantir starting in 2021.
    • FDE space is considered the single best way to win in enterprise software.
  • Examples:
    • Kevin speaking at conferences about FDE and helping companies set up FDE functions.
  • Terminology:
    • Forward Deployed Engineer (FDE)
    • Enterprise software
  • Why it matters: Establishes credibility and context for discussing the FDE role across top-tier tech companies.

What is an FDE and Market Trends

  • Explanation: Discussion on why FDE is on the rise, how it bridges the gap between technical products and non-technical customers, and dispelling hype versus reality.
  • Key claims:
    • FDE is not sales engineering or traditional software engineering; it bridges technical solutions with non-technical customers.
    • Agentic AI makes it easier to build powerful solutions, but also makes them more complicated for traditional customers.
    • Lack of technical talent on the customer's payroll should not be a barrier to startup success.
  • Examples:
    • Comparing technical products like Temporal API or GitHub with user-facing tools like Slack or Jira.
    • Palantir Foundry and app-building platform serving Fortune 500 non-tech companies like CPG, manufacturing, and aviation.
  • Terminology:
    • Ideal Customer Profile (ICP)
    • Agentic AI
    • LLM
  • Why it matters: Helps candidates and founders determine if an FDE function is actually needed versus standard sales engineering or standard software engineering.

The Three Hats of an FDE

  • Explanation: Kevin breaks down the FDE role into three core capabilities: consultant, product manager, and software engineer.
  • Key claims:
    • The first part of any engagement is listening and gaining trust from the customer.
    • The end output of an FDE engagement must be software (invention of SKUs).
    • FDE is an engineering capacity and role, crossing over with professional services or consulting practices.
  • Examples:
    • Listening to executive sponsors, VPs, and CEOs to uncover problems that keep them up at night.
  • Terminology:
    • Consultant
    • Product Manager
    • Software Engineer
    • SKUs
  • Why it matters: Provides a clear framework for understanding what an FDE actually does day-to-day.

Discovery, Listening, and Scoping Trade-offs

  • Explanation: Explaining how FDEs handle discovery, latency issues, customer symptoms vs. root problems, and scope management.
  • Key claims:
    • The problem is rarely the problem; customers usually tell you the symptom.
    • FDEs must separate signal from noise, aiming for highest impact relative to easiest problem to solve.
    • Iterative development and starting with the smallest scope that yields net increase in conversion is critical.
  • Examples:
    • A sales dashboard loading slowly where the customer cites latency as the issue, but the actual business goal is sales conversion.
  • Terminology:
    • Latency
    • Conversion rate
    • Signal from noise
    • Iterative development
  • Why it matters: Teaches the practical execution of FDE client engagements without boiling the ocean.

Interviewing and Recruiting for FDE Roles

  • Explanation: Discussing who is the right fit for an FDE role, what skills to brush up on, and how to prepare for interviews.
  • Key claims:
    • FDE sits at the intersection of consulting, product, and engineering.
    • Candidates need clear communication, executive presence, and strong judgment.
    • FDE interviews require demonstrating end-to-end ownership of customer problems, not just writing code.
  • Examples:
    • Preparing for FDE interviews by studying great products and understanding how they are made.
  • Terminology:
    • Executive presence
    • Completeness
    • End-to-end ownership
    • Test cases
  • Why it matters: Provides actionable guidance for job seekers wanting to break into Forward Deployed Engineering.

Key Points

FDE is not Sales Engineering or Software Engineering

  • Explanation: Sales engineering is pre-sales focused on demos to win deals. Software engineering builds one-to-many products scaling to thousands of users. FDE owns everything from discovery to delivery end-to-end, building one-to-one or highly tailored solutions for specific high-value customers.
  • Evidence: Kevin Bai's breakdown of pre-sales vs. end-to-end ownership.
  • Practical implication: Candidates must understand they carry the responsibility of delivering software that solves real customer problems end-to-end.

The Three Hats: Consultant, Product Manager, Software Engineer

  • Explanation: An FDE must wear three hats simultaneously: listening like a consultant, prioritizing scope like a product manager, and building software like an engineer.
  • Evidence: Detailed breakdown of the FDE function during customer deployments.
  • Practical implication: Successful FDEs cannot rely purely on coding skills; communication, empathy, and product sense are equally vital.

Listening is the Primary Communication Skill

  • Explanation: The main communication skill of an FDE is active listening—getting customers to share the real problems that keep them up at night, not just surface-level symptoms.
  • Evidence: Mentorship experience sharing how customers state symptoms while FDEs uncover root causes.
  • Practical implication: FDEs must practice active listening and build deep trust with executive sponsors.

Iterative Scope over Perfection

  • Explanation: Trying to build a perfect solution in one go leads to failed pilots. FDEs look for the smallest scope that drives net business value and iterate rapidly.
  • Evidence: Discussion around pilot timelines (e.g., 2-4 week pilots).
  • Practical implication: Ship a functional first version and iterate based on customer feedback rather than waiting for perfection.

Frameworks, Models & Processes

The Three Hats of FDE

  • How it works: An FDE operates effectively by balancing three distinct professional roles during an engagement.
  • Components:
    • Consultant (listening, gaining trust, uncovering root problems)
    • Product Manager (scoping, prioritizing high-impact wins, managing trade-offs)
    • Software Engineer (building, testing, and delivering the actual software solution end-to-end)
  • When to use: Evaluating candidates for FDE roles or structuring an FDE team within an enterprise software company.

Examples & Case Studies

A customer complains that a sales dashboard is too slow and has latency issues.

  • Illustrates: Customers talking about symptoms rather than root business problems.
  • Lesson: FDEs must dig deeper to understand whether latency actually impacts sales conversion or if the underlying business problem is different.

An FDE builds a function with strict integer inputs, but a customer passes strings, breaking the app.

  • Illustrates: The importance of completeness and anticipating edge cases and misuse.
  • Lesson: FDEs must design software with robust instructions and safeguards for non-technical users.

Actionable Takeaways

  • Immediate:
    • Familiarize yourself with the three hats of FDE: consultant, product manager, and software engineer.
    • Practice active listening to uncover root customer problems rather than surface symptoms.
    • Focus on iterative development and small scope for pilot projects.
  • Strategic:
    • FDE is the single best way to win in enterprise software when dealing with non-technical ICPs.
    • Build executive presence and the ability to present technical solutions clearly to senior leadership.
    • Evaluate candidates based on ownership of the problem rather than just project execution.
  • Questions to investigate:
    • Does our product require an FDE function, or is standard sales engineering sufficient?
    • How do we balance delivery speed with software quality and completeness during customer pilots?

Claims Worth Verifying

  • FDE is the single best way to win in enterprise software. (opinion)
  • Palantir has existed for over 20 years and pioneered the FDE model. (fact)

Notable Quotes

"The main communication skill of an FDE is listening." (at 122:47) "Fail to prepare, prepare to fail." (at 440:56)

Compressed Summary

  • FDE bridges technical products and non-technical enterprise customers.
  • The role combines consulting, product management, and software engineering.
  • Active listening and unearthing root customer problems are paramount.
  • Iterative scoping and end-to-end ownership drive pilot success.
  • Keywords: forward deployed engineer, enterprise software, palantir, rippling, consulting
  • Core insight: Forward Deployed Engineering is an end-to-end customer-facing software engineering role that combines consulting, product management, and technical delivery to solve complex problems for non-technical enterprise clients.

Core insights

6
Predictionhigh noveltymoderate evidence

Agentic AI shifts the difficulty frontier from building powerful solutions to making them comprehensible and deployable for traditional, non-technical customers; FDE-style forward deployment becomes the critical differentiation.

Why it matters

It redirects engineering investment from raw agent capability toward deployment, integration, and customer-success tooling—the absence of which becomes the bottleneck.

Generalization

As AI lowers build cost, the moat of an AI product shifts to the adaptation layer that closes the gap between generic capability and specific customer workflow.

Agentic AI makes it easier to build powerful solutions, but also makes them more complicated for traditional customers.
Open source video
Mental Modelmedium noveltystrong evidence

FDE is not pre-sales demos or one-to-many product engineering; it is a distinct role with end-to-end ownership from discovery to delivery for one-to-one or highly tailored customer solutions.

Why it matters

If you are building an AI system for enterprise, this model says to put an accountable owner in front of each high-value customer rather than splitting responsibilities across sales, product, and engineering.

Generalization

Architectures should include a customer-specific adaptation and ownership layer, not just a platform, to avoid context loss in handoffs.

Sales engineering is pre-sales focused on demos to win deals. Software engineering builds one-to-many products scaling to thousands of users. FDE owns everything from discovery to delivery end-to-end, building one-to-one or highly tailored solutions for specific high-value customers.
Open source video
Mechanismmedium noveltystrong evidence

The stated problem is usually a symptom; effective forward deployment starts by listening to executives to find root problems, then separates signal from noise before scoping.

Why it matters

Agentic workflows that naively execute user requests can optimize for the wrong objective; they need a discovery and reframing phase that maps symptoms to actual business outcomes.

Generalization

Task-taking agents should include a root-cause check before committing to the literal instruction, especially in high-stakes enterprise contexts.

The problem is rarely the problem; customers usually tell you the symptom.
Open source video
Listening to executive sponsors, VPs, and CEOs to uncover problems that keep them up at night.
Open source video
FDEs must separate signal from noise, aiming for highest impact relative to easiest problem to solve.
Open source video
Practicemedium noveltystrong evidence

Forward-deployed delivery is intentionally iterative: ship the smallest scope that creates net business value, then iterate, because trying to build a perfect solution in one go leads to failed pilots.

Why it matters

For AI deployments, a 2-4 week pilot with a narrowly scoped measurable outcome is more reliable than a large perfect rollout; iteration over business metrics prevents failed pilots.

Generalization

Use smallest-viable-deployment patterns and business-outcome metrics rather than full-feature launches.

FDEs look for the smallest scope that drives net business value and iterate rapidly.
Open source video
Trying to build a perfect solution in one go leads to failed pilots.
Open source video
Ship a functional first version and iterate based on customer feedback rather than waiting for perfection.
Open source video
Practicehigh noveltymoderate evidence

Active listening and executive presence are primary FDE skills; trust with executive sponsors is a precondition for uncovering real problems and deploying software.

Why it matters

AI engineering teams often focus on code and model capability while underinvesting in stakeholder trust; without executive-level trust, the AI system may never get access, data, or adoption.

Generalization

For enterprise AI, treat listening as a specification-gathering capability and build trust-focused engagement practices, not as a soft extra.

The first part of any engagement is listening and gaining trust from the customer.
Open source video
The main communication skill of an FDE is active listening—getting customers to share the real problems that keep them up at night, not just surface-level symptoms.
Open source video
Practicemedium noveltystrong evidence

The output of an FDE engagement must be actual software—'invention of SKUs'—not recommendations; value is realized only when bespoke software enters the customer's operations.

Why it matters

AI teams should not stop at demos or advisory reports; the deployment must create a deployed, owned software artifact tied to the customer's operation.

Generalization

Measure AI value by deployed and integrated software artifacts with named owners, not by capability demonstrations or evaluation-only results.

The end output of an FDE engagement must be software (invention of SKUs).
Open source video

Deep dives

5

Productizing the Forward-Deployed Layer

Research question

How can recurring deployment patterns from FDE engagements be systematically harvested into reusable platform abstractions without premature generalization?

Why

FDE one-to-one delivery does not scale economically to many customers; if AI products rely on forward deployment, we need a method to turn bespoke solutions into product capabilities.

FDE owns everything from discovery to delivery end-to-end, building one-to-one or highly tailored solutions for specific high-value customers.
Open source video
Agentic AI makes it easier to build powerful solutions, but also makes them more complicated for traditional customers.
Open source video
Source video

Evaluating AI Agents for FDE Discovery

Research question

Can a conversational AI agent perform the listening and root-cause-discovery function of an FDE, and what evaluation methodology measures whether it uncovers root problems instead of acting on symptoms?

Why

If agents can replicate discovery, FDE effort can focus on exception handling; if not, human-in-the-loop remains necessary. This defines the division of labor between agents and FDEs in enterprise AI.

The problem is rarely the problem; customers usually tell you the symptom.
Open source video
The first part of any engagement is listening and gaining trust from the customer.
Open source video
Source video

Operationalizing Smallest-Scope Delivery for AI Pilots

Research question

What is the operational definition of 'smallest scope that yields a net increase in the target business metric', and how can teams determine it under uncertainty before deployment?

Why

Teams know to iterate, but without a method for choosing the minimal viable scope, pilots either overbuild or under-deliver. A formal operationalization would improve AI deployment success rates.

FDEs look for the smallest scope that drives net business value and iterate rapidly.
Open source video
Trying to build a perfect solution in one go leads to failed pilots.
Open source video
Source video

Agent Architectures with a Discovery Loop

Research question

What architectural components are needed in an AI agent system to perform root-cause reframing before executing a customer request, and how do they affect task outcomes?

Why

Agents that naively execute user requests risk optimizing for the wrong objective in high-stakes enterprise contexts; a discovery loop could be the difference between failed pilots and real business impact.

The problem is rarely the problem; customers usually tell you the symptom.
Open source video
FDEs must separate signal from noise, aiming for highest impact relative to easiest problem to solve.
Open source video
Source video

The Future of Forward-Deployed Engineering in the Age of Agentic AI

Research question

Will increasing agentic AI capability amplify the need for human FDEs (as adoption complexity grows) or automate the FDE function away?

Why

This determines how AI companies allocate talent and product investment; it is a strategic question for any company selling to non-technical enterprise customers.

Agentic AI makes it easier to build powerful solutions, but also makes them more complicated for traditional customers.
Open source video
Lack of technical talent on the customer's payroll should not be a barrier to startup success.
Open source video
Source video

Article ideas

4

The Real Moat for AI Products Is the Forward-Deployed Layer

As agentic AI makes building powerful solutions easier, the differentiator for enterprise AI shifts from model capability to the deployment and adaptation layer that makes the solution comprehensible and usable by non-technical customers.

Angle

Argue that AI startups should invest in FDE-style capacity and deployment tooling as a core product feature, not a support afterthought.

Source video

Stop Optimizing for User Instructions: Give Agents a Discovery Loop

Enterprise AI agents should not treat explicit user requests as ground truth; they need a discovery phase that separates symptoms from root problems to avoid optimizing for the wrong business outcome.

Angle

Challenge the common design of instruction-following agents, using FDE discovery practice as evidence for a pattern.

Source video

Why Your Enterprise AI Pilot Fails: You Aimed for Perfect Instead of the Smallest Valuable Scope

The leading cause of failed enterprise AI pilots is over-scoping the initial deployment; shipping a minimal but measurable version and iterating based on business metrics is a more reliable pattern.

Angle

Use FDE iterative delivery principles to reframe AI pilot design; include a concrete scoping heuristic.

Source video

The Case for End-to-End Ownership in AI Systems: Context Loss Kills Adoption

AI companies serving enterprise customers should assign an accountable forward-deployed owner to high-value accounts from discovery to delivery, rather than splitting handoffs across sales, product, and support, because context loss in handoffs is the real adoption killer.

Angle

Contrast FDE model with traditional role separation; argue for account-level technical owners in AI system architecture.

Source video

Project ideas

4

Agent Discovery Evaluator

beyond-evals

A dialogue agent with a discovery loop can identify the underlying business problem from a symptomatic support request at a level comparable to an FDE, measured by expert agreement on root-problem extraction.

Proof of concept

Build a benchmark set of enterprise customer requests paired with root problems and business metrics; implement a discovery agent that asks clarifying questions before committing to a task; run it against the benchmark and collect expert ratings on reframed problem statements.

Measurement

Expert precision/recall of root-problem identification vs FDE baseline; quality of clarifying questions; downstream metric lift in a simulated deployment.

Source video

FDE Reuse Miner

new

Pattern-mining across historical one-to-one FDE deployments will reveal recurring components that can be extracted into reusable modules, reducing the time-to-deployment for new engagements by at least 30% without decreasing net business value.

Proof of concept

Instrument a lightweight logging layer in a sample of FDE-style deployments; after 10 engagements, apply clustering to extract common solution patterns; build a module library; measure reuse in next 5 engagements.

Measurement

Percentage of reused code/config vs greenfield; median deployment time; net business value from before/after.

Source video

Smallest Effective Pilot Harness

new

Using a scoping advisor that recommends the smallest feature set predicted to move a target business metric will reduce failed enterprise AI pilots by 30% compared to a full-scope launch, as measured by the proportion of pilots achieving net business value within four weeks.

Proof of concept

Create a scoping tool that takes a business goal and available capabilities, uses historical pilot data and approximate causal impact weights to propose a minimal configuration; run a controlled simulation on synthetic customer scenarios.

Measurement

Pilot success rate (net business value in 4 weeks), time-to-value, number of iterations needed.

Source video

Forward-Deployed Trust Tracker

new

Executive sponsor trust scores, measured through structured engagement signals in the first two weeks of an FDE engagement, are a leading indicator of deployment success and can predict pilot failure with at least 75% accuracy.

Proof of concept

Develop a short trust probe and activity log (meeting attendance, response latency, willingness to share data) for FDE engagements; collect data over several pilots and correlate with final deployment outcome.

Measurement

Predictive accuracy of pilot failure/success at week 2; correlation coefficient between trust score and final net business value.

Source video

Architectural implications

4

Agentic AI changes where value is created: solutions become easier to build, but customer-side complexity rises.

Before

Products were differentiated primarily by internal capability and assumed technical buyers or an in-house engineering team.

After

Products need a forward-deployed adaptation layer that configures, explains, and integrates the agent for non-technical customers.

Consequence

Build platforms with customer-specific extension points, configuration surfaces, and deployment tooling; treat FDE capacity as part of the product architecture.

Source video

FDE has end-to-end ownership from discovery to delivery for specific high-value accounts.

Before

Sales engineering handled demos, product engineering built one-to-many features, and consulting delivered advice.

After

A single role owns the full lifecycle of a customer's deployed solution.

Consequence

Agent-system teams should assign account-level technical owners and maintain persistent customer context rather than relying on sequential handoffs.

Source video

FDE separates symptom from root problem, aiming for highest-impact/easiest-scope combinations.

Before

Requirements are taken at face value and immediately translated into features.

After

Requirements pass through discovery and root-cause reframing before scope is selected.

Consequence

Agent architectures need a discovery loop that can ask questions, challenge stated goals, and use business metrics to prioritize actions.

Source video

Iterative small-scope pilots outperform one-shot perfection.

Before

Rigorous full-feature development before release.

After

Release a minimal pilot with net business value, then iterate rapidly.

Consequence

Support progressive deployment, scope flags, and business-outcome telemetry for AI agents.

Source video

Tradeoffs and failure modes

4

End-to-end tailored delivery vs product scale

Benefit

Solutions fit high-value customers and avoid handoff loss.

Cost or risk

One-to-one engineering does not scale economically to many customers.

FDE owns everything from discovery to delivery end-to-end, building one-to-one or highly tailored solutions for specific high-value customers.
Open source video
Source video

Agentic AI capability vs customer complexity

Benefit

Powerful solutions are faster to build.

Cost or risk

Traditional customers struggle to adopt and operate them.

Agentic AI makes it easier to build powerful solutions, but also makes them more complicated for traditional customers.
Open source video
Source video

Iterative small scope vs perfect final solution

Benefit

Faster value delivery, real feedback, fewer failed pilots.

Cost or risk

Initial version may lack completeness; requires continued iteration.

Trying to build a perfect solution in one go leads to failed pilots.
Open source video
Source video

Acting on stated symptoms vs discovering root problems

Benefit

Fast response to the explicit request.

Cost or risk

Risk optimizing the wrong thing and wasting effort.

The problem is rarely the problem; customers usually tell you the symptom.
Open source video
Source video

Open questions

4

How do you productize a forward-deployed layer so one-to-one customer solutions become repeatable platform capabilities instead of permanent bespoke code?

Why unresolved

The summary presents FDE as one-to-one and tailored, but does not explain how learnings are reused across customers.

Research direction

Track recurring deployment patterns, measure reuse, and gate platform abstractions behind observed customer demand.

Source video

Can an AI agent perform the listening and root-cause-discovery part of FDE, and how would we evaluate that?

Why unresolved

Listening is described as a human skill in the summary; no evaluation method is given.

Research direction

Build a discovery agent that interviews stakeholders, compare its problem formulation to the literal symptom, and measure downstream outcome improvement.

Source video

What is the operational definition of 'smallest scope that yields net increase in conversion'?

Why unresolved

The summary states the principle but not a method for choosing scope under uncertainty.

Research direction

A/B test varying pilot scopes to find the minimum viable deployment that moves the target business metric.

Source video

If agentic AI increases complexity for traditional customers, will the FDE function become more critical or eventually be automated away?

Why unresolved

Two opposing pressures exist: agent capabilities may automate integration, but adoption complexity may create a greater need for human bridging.

Research direction

Compare adoption and success rates for agentic products with and without forward-deployed support over time.

Source video

Key claims

8
comparativeVerification needed

FDE is not sales engineering or traditional software engineering; it owns discovery to delivery end-to-end for one-to-one or highly tailored solutions.

Evidence

Sales engineering is pre-sales focused on demos to win deals. Software engineering builds one-to-many products scaling to thousands of users. FDE owns everything from discovery to delivery end-to-end, building one-to-one or highly tailored solutions for specific high-value customers.

Question

Do real FDE job descriptions match this definition, or do many companies use the title for a hybrid of sales and support?

Source video
opinionVerification needed

FDE space is considered the single best way to win in enterprise software.

Evidence

FDE space is considered the single best way to win in enterprise software.

Question

Is there causal evidence linking FDE presence to enterprise deal win rates?

Source video
causalVerification needed

Agentic AI makes it easier to build powerful solutions, but also makes them more complicated for traditional customers.

Evidence

Agentic AI makes it easier to build powerful solutions, but also makes them more complicated for traditional customers.

Question

What empirical evidence supports the claim that agentic AI adoption is more difficult for traditional customers?

Source video
opinionVerification needed

The problem is rarely the problem; customers usually tell you the symptom.

Evidence

The problem is rarely the problem; customers usually tell you the symptom.

Question

In controlled discovery studies, how often does the stated request diverge from the underlying business problem?

Source video
opinionVerification needed

FDEs look for the smallest scope that drives net business value and iterate rapidly.

Evidence

FDEs look for the smallest scope that drives net business value and iterate rapidly.

Question

Does smallest-scope-first delivery outperform full-scope delivery in measured business outcomes?

Source video
opinionVerification not requested

An FDE must wear three hats simultaneously: listening like a consultant, prioritizing scope like a product manager, and building software like an engineer.

Evidence

An FDE must wear three hats simultaneously: listening like a consultant, prioritizing scope like a product manager, and building software like an engineer.

Source video
predictionVerification needed

Lack of technical talent on the customer's payroll should not be a barrier to startup success.

Evidence

Lack of technical talent on the customer's payroll should not be a barrier to startup success.

Question

Do startups with FDE-style deployment layers succeed with non-technical customers more often than those without?

Source video
opinionVerification needed

FDE interviews require demonstrating end-to-end ownership of customer problems, not just writing code.

Evidence

FDE interviews require demonstrating end-to-end ownership of customer problems, not just writing code.

Question

What specific interview tasks are used to assess end-to-end ownership across FDE hiring pipelines?

Source video

Connections

5