Reqflow
← All tracks
FDE Track

Forward Deployed Engineer

The role Palantir invented and AI infrastructure companies revived: the engineer whose workspace is the customer's environment, not the company's product backlog. 11 lessons, ~122 minutes of reading.

After this track you will be able to

Operate as a forward-deployed engineer: frame a customer's real problem, ship a working demo in 48 hours, navigate deployment into their environment, and turn a one-off engagement into productionized work without losing the room politically.

Interview approach for FDE roles
  1. 1.Lead with the customer, not the architecture. FDE interviews are usually scenarios - start by interrogating the situation, not by drawing boxes.
  2. 2.Name the framing question you'd ask first ('walk me through the last time this came up') before you propose anything.
  3. 3.Pick a 48-hour vertical slice and describe it explicitly - one workflow, end to end, what's real, what's stubbed.
  4. 4.Surface the integration boundaries (data in, data out, auth) early. Strong candidates ask about these before discussing internal architecture.
  5. 5.Talk about evals like you'd talk about tests, especially for AI-flavored scenarios. Customer-specific data is the eval set.
  6. 6.Address the non-technical stakeholders. Who's the executive sponsor, the manager, the end user, and where do their definitions of success conflict?
  7. 7.Close with a plausible path from 'demo works' to 'in production' - security review, UAT, runbook, handoff. The FDE who skips this loses months at the customer.

Lessons

11 total
  1. 1
    What an FDE Actually Does
    Foundational11 min

    The role was invented at Palantir in 2006. Twenty years later it is the most-discussed engineering title at OpenAI and Anthropic. Here is what it actually is.

    →
  2. 2
    Problem Framing in the Customer's Environment
    Foundational11 min

    Translating what the customer asked for into what they actually need, without becoming the person who says 'actually you don't need that.'

    →
  3. 3
    The 48-Hour Demo: Prototyping for a Customer
    Applied12 min

    OpenAI structures every FDE engagement around an early-scoping phase measured in days, not weeks. Here is what that compresses into and why speed is the actual deliverable.

    →
  4. 4
    Thin Vertical Slices: Full-Stack Pragmatism
    Applied11 min

    Why one feature shipped end-to-end beats six features built horizontally, and how Palantir's 'Delta' model formalizes the idea.

    →
  5. 5
    Deploying Into Customer Environments
    Advanced12 min

    On-prem, air-gapped, their cloud, their data. What changes when your code has to run in someone else's house, and why Anthropic now ships MCP servers as the unit of deployment.

    →
  6. 6
    FDE in the AI Era
    Advanced13 min

    Foundation models revived a role Palantir invented in 2006. In 2025-2026, OpenAI runs a $4B deployment JV, Anthropic launched a $1.5B FDE consulting JV, and Ramp organizes FDEs into pods. Here is what changed.

    →
  7. 7
    Productionizing Custom Work
    Advanced11 min

    When does Palantir push a Delta's work into the Foundry platform? When does it stay bespoke? The rule of three, the handoff problem, and the contract you owe the team that inherits the code.

    →
  8. 8
    Communicating With Non-Engineers
    Applied10 min

    Demos as the primary language, the three-sentence executive readout, and how to say no without losing the room.

    →
  9. 9
    Preparing for the FDE Interview
    Applied10 min

    What's defensibly known about FDE interview loops, what we deliberately won't make up, and how to prep for the parts that are actually documented.

    →
  10. 10
    Comp and Where FDEs Go Next
    Applied11 min

    Verified Palantir comp from Levels.fyi, a frank acknowledgment of what non-Palantir comp data is reliable (very little), and three named ex-FDE founders with real funding totals.

    →
  11. 11
    Failure Modes and the Honest Case Against
    Advanced10 min

    The org-level trap that turns FDE teams into consulting shops, the misappropriation problem the role has on LinkedIn, and the personal-burnout dimension we deliberately won't fake.

    →

Reference

Glossary →

FDE vocabulary: Delta, MCP server, sub-agent, trust ladder, more.

Reading list →

Primary sources, ex-FDE essays, current job listings.

Other tracks

System Design →

Interactive architecture case studies.

All tracks →

See everything Reqflow covers.