Forward Deployed Engineer
The role that lives between sales engineering and product engineering: scope problems, ship POCs fast, and turn pilots into production.
An FDE shows up at a customer with an idea and a laptop and leaves with a working integration. The skill mix is unusual — deep technical fluency, customer empathy, and the ability to ship under ambiguity.
The FDE Mindset
Why this role exists, and what makes it different from sales engineering or product.
- 01
What an FDE Actually Does
coreCustomer-facing, code-shipping, problem-scoping. You own the demo and the integration.
- 02
Customer Empathy
coreThe real problem is rarely the stated problem. Learn to listen past the ask.
- 03
Operating in Ambiguity
coreSpecs will be incomplete. Ship a thin slice, then iterate with the customer in the room.
SDK Fluency
Be the person who can wire up any model API in 20 minutes.
- 01
Anthropic SDK
coreMessages API, tool use, prompt caching, streaming. Know it cold.
- 02
OpenAI SDK
coreResponses API, function calling, structured outputs. The other half of the market.
- 03
Gemini SDK
recommendedLong-context workloads, multimodal. Worth fluency even if it's not your default.
- 04
Streaming & Realtime
recommendedSSE, websockets, partial JSON parsing — what makes demos feel alive.
Solution Discovery
Before you write code, get the right problem on the table.
- 01
Technical Discovery
coreArchitecture diagrams, data shape, auth model, deployment constraints. Ask early.
- 02
Scoping a POC
corePick the smallest thing that earns the next meeting. Cut features ruthlessly.
- 03
Defining Success
coreWrite down what 'it works' means before you build. Get the customer to agree.
Demo Engineering
Build POCs that produce wow moments — and survive a live customer call.
- 01
Fast POCs with an Agent
coreUse vibe-coding tools (Claude Code, Codex) to compress a week of work into a day.
- 02
Demos That Don't Crash
coreMock the unreliable parts. Cache the slow calls. Have a backup plan for offline.
- 03
Realistic Sample Data
recommendedGeneric data feels generic. Use the customer's own data shape (or a close synth).
- 04
Narrating a Live Demo
recommendedTelegraph what's about to happen. Pause for reactions. Don't read the screen.
Integration Patterns
From POC to working integration with the customer's existing stack.
- 01
Auth Patterns
coreOAuth, SAML, service accounts, API keys. The boring stuff that breaks every integration.
- 02
Webhooks & Async
coreLong-running AI calls don't fit in a 30-second HTTP request. Design for async.
- 03
Data Residency & PII
recommendedWhere the data lives, what's logged, what's redacted. Enterprise asks early.
- 04
Customer-Facing Observability
recommendedGive the customer a dashboard. Trust comes from visibility.
Production Handoff
Turn the pilot into something the customer's team can own.
- 01
Runbooks & Docs
coreWrite the docs you wish you'd had. Future-you and the customer's ops team thank you.
- 02
Handoff Criteria
coreDefine what 'production-ready' means. Don't hand off a prototype with a bow on it.
- 03
Success Metrics in Production
recommendedSet up the dashboards that prove the value six months in. Renewal depends on it.
Soft Skills
The career multipliers nobody puts on the JD.
- 01
Written & Spoken Comms
coreCrisp updates, executive summaries, technical explanations for non-technical audiences.
- 02
Leading Without Authority
recommendedYou're not the customer's manager. Influence through expertise and reliability.
- 03
Domain Curiosity
coreLearn the customer's business. The technical answer always lives inside the business context.