On-Site vs Remote FDE: Travel Realities, Comp, and What to Expect

Most engineers interviewing for Forward Deployed roles fixate on the travel. Recruiters often give a vague “25–50%” number that feels abstract until you’re staring at a Monday 6 AM boarding pass with a half-built integration running on your laptop.

This piece breaks down the real difference between on-site and remote FDE work—not just logistics, but the engineering decisions, compensation mechanics, and trust-building patterns that define each mode.

The Two Modes: Embedded vs. Remote-First

An FDE operates in two distinct modes depending on the engagement phase and customer sensitivity.

Mode 1: Embedded (On-Site) You’re physically inside the customer’s environment—often a classified facility, a trading floor, or a manufacturing plant. You’re debugging against their actual data, navigating their proxy servers, and whiteboarding with their architects. Travel tends to be bursty: 3-5 days on-site, then 1-2 weeks remote for build-out.

Mode 2: Remote-First You’re shipping from your home office, connecting via secure tunnels or customer-provided VPNs. This mode works for SaaS-heavy integrations, API-layer work, and post-POC scaling phases. You’re still customer-facing daily, just through Slack shared channels, Zoom, and async video walkthroughs.

The reality: most FDEs live in a hybrid. The question isn’t “do you travel?” but “do you control the travel cadence?” Senior FDEs negotiate this hard.

Compensation and the Travel Premium

On-site work carries a real premium. Companies aren’t just paying for your skills—they’re paying for the physical exhaustion, context-switching cost, and the fact that you can’t interview elsewhere while on a client site.

Travel-heavy FDEs often accumulate airline and hotel status within 6 months. This isn’t a perk—it’s a tool. Status means same-day flight changes when a POC breaks, lounge access for calls between meetings, and suite upgrades that let you work without lobby noise. Treat points as part of comp; a heavy-travel FDE can extract $10K–$15K/year in untaxed value from points alone.

The Weekly Rhythm: On-Site Rituals That Justify the Flight

Flying across the country to sit on Zoom defeats the purpose. Effective on-site FDEs run a specific set of rituals that remote work can’t replicate. For a deeper dive into the trust-building mechanics behind these rituals, see the Palantir-style FDE embed playbook.

Monday Morning: War Room Setup You’re in a conference room with the customer’s lead engineer, a whiteboard, and access logs. The goal: reproduce the error that’s been blocking the pipeline since Friday. You don’t leave until it’s root-caused. This builds credibility faster than any slide deck.

Tuesday–Wednesday: Pairing on Their Stack You write code side-by-side with their team. Not demos—actual commits against their fork. You’re navigating their CI/CD, their linter rules, their bizarre internal library that wraps fetch with a custom auth header. Remote pairing tools exist, but nothing matches the bandwidth of two engineers pointing at the same monitor.

Thursday: The Executive Readout A 15-minute standup with the customer’s VP of Engineering. You show a live dashboard, not a PowerPoint. You flag risks without sugarcoating. This is where renewals are won or lost.

Friday: Handoff and Flight Back You document everything in a running log, open PRs against both your codebase and theirs, and set the remote work plan for the next two weeks.

The Remote Toolchain: Async Demos and Secure Tunnels

Remote FDE work lives or dies on tooling. You’re not there to tap someone on the shoulder, so your async artifacts must be bulletproof.

The Async Demo Pattern Instead of scheduling a live call to show a working feature, record a 3-minute Loom video. Walk through the code change, show the output, and highlight one edge case you handled. Post it in the shared Slack channel. The customer engineer watches it on their own time, leaves timestamped comments, and you iterate. This pattern compresses what used to be a 30-minute synchronous meeting into 10 minutes of total attention.

Secure Access Without IT Tickets Enterprise customers hate opening firewall ports. Modern FDEs use Tailscale or WireGuard-based overlay networks that don’t require inbound rules. You ship a small agent that initiates an outbound connection to your coordination server. Once the tunnel is up, you can hit their staging APIs, query their test databases, and debug integrations without a 3-week infosec review.

Shared Issue Tracking Never track work in your company’s internal Jira and expect the customer to trust you. Mirror relevant tasks into a shared Linear board or a GitHub Project that both teams can see. Transparency in task status prevents the “what is the FDE even doing?” conversation that kills engagements.

The Trust Handoff: Why the First Week Must Be In-Person

There’s a specific sequence that determines whether a remote engagement succeeds. It starts with a physical handshake.

Week 1 must be on-site. This is non-negotiable for complex integrations. During those five days, you establish:

• The shared vocabulary: their team calls it a “data plane,” you call it a “pipeline.” You align on terminology by drawing architecture on the same whiteboard.
• The escalation path: you learn who actually unblocks things versus who’s on the org chart. This social graph is invisible on Zoom.
• The real constraints: the “air-gapped” environment that actually has a USB port; the “deprecated” endpoint that still serves 80% of traffic.

After Week 1, you can operate remotely for 4-6 weeks before needing another on-site pulse check. The ratio that works for most FDEs: 1 week on-site per 4-6 weeks remote. This translates to roughly 20-25% travel, not the 50% recruiters sometimes quote.

When you do go remote, the handoff to core engineering becomes critical. As you stabilize the integration, you’ll need to scale yourself by handing off to core engineering. This isn’t a failure—it’s the design pattern. The FDE proves the integration works, documents the edge cases, then core engineering productizes it while you move to the next account.

FAQ: On-Site vs Remote FDE

Q: Can a forward deployed engineer work 100% remote? Yes, but it’s rare and usually limited to SaaS-only products with no on-prem component. If your product involves deploying to a customer’s VPC, on-prem server, or classified environment, expect at least quarterly on-site visits. Pure remote FDE roles do exist at companies like GitLab or Zapier, but they’re the exception.

Q: How do I negotiate travel percentage? Don’t negotiate the percentage—negotiate the cadence. “One week per month” is clearer than “25%” and prevents death-by-a-thousand-day-trips. Also negotiate who owns the schedule: you want veto power over trips that conflict with personal commitments.

Q: What’s the biggest remote FDE failure mode? The “black box” problem. The customer stops hearing from you except in weekly status calls, and they assume nothing is happening. Combat this with public artifacts: daily Slack summaries, Loom demos of in-progress work, and a shared dashboard showing integration health. For a concrete example of building such a dashboard, check out this on-call incident summarizer that drafts postmortems—the same pattern applies to integration monitoring.

Q: Do on-site FDEs get paid more? Yes, typically a $10K–$20K base salary premium plus untaxed travel perks. But the real comp advantage is career velocity: on-site FDEs build executive relationships that accelerate promotions and open paths to solutions architecture or customer success leadership.

Q: How do I stay productive on travel days? Travel days are lossy. Don’t pretend otherwise. Use flights for deep work that doesn’t require internet: writing design docs, reviewing PRs you’ve cached locally, or recording code walkthroughs. Airport lounges are your office—invest in a credit card that grants access. Noise-canceling headphones are not optional.