How Ramp Does Forward Deployed Engineering

These are my notes from the Forward Deployed Engineering workshop at @aiDotEngineer on June 30, 2026.

Session: How Forward Deployed Engineering Is Done at @RampLabs
Speaker: Leo Mehr
Company: Ramp
X: @LeoMehr

Role

Ramp's FDE function lives inside engineering.

It focuses on helping Ramp win with large enterprise customers and covers:

• Deployed developer APIs.
• A new AI services business.

It is a distinct function, not "boss mode technical sales."

Principle 1: Always Be Scoping

The wrong reflex is to say yes and build fast. FDEs should interrogate urgency and context before committing.

Questions include:

• Who uses this integration?
• Have workarounds been exhausted?
• Can the customer self-serve?
• Do other prospects or customers have the same issue?

Validate basic assumptions before building.

Principle 2: Scale the Tokens

FDEs should use AI to scale:

• Context gathering.
• Scoping.
• Candidate identification.

Ramp uses a Notion agent in #FDE-requests to ask submitters multi-round scoping questions. This reduced latency from hours or days to seconds and saved about 20% of scoping time.

Future FDEs may spend more time managing agentic pipelines while retaining judgment over quality.

My Take

Ramp's FDE principle is a useful correction to the default builder instinct. The wrong reflex is to say yes and ship fast before understanding whether the work is urgent, generalizable, or even necessary.

The practical takeaway is that FDE teams should scale scoping as aggressively as they scale building. Ramp's Notion agent in #FDE-requests is a good example: use AI to gather context and pressure-test assumptions, while keeping humans responsible for judgment.