Forward Deployed Engineer is one of the fastest-growing, highest-paid technical roles in AI — and unlike many senior titles, there's no single degree or certification that gets you in. That's good news: the path is open to strong engineers from several backgrounds. This roadmap lays out exactly how to get there.
First, is FDE right for you?
FDE is a great fit if you:
- Enjoy solving open-ended technical problems end to end
- Like working directly with customers (not behind a ticket queue)
- Are energized by variety — a new business problem every few weeks
- Can stay calm when something breaks in production in front of a customer
It's a poor fit if you want to go deep on one system for years, or if customer-facing work drains you. Be honest with yourself here; the role lives or dies on customer interaction.
The four skill pillars
Every FDE needs competence across four areas. You don't need to be world-class at all of them — but you can't be missing one.
- Full-stack engineering — you own implementations end to end: backend, APIs, a bit of frontend, deployment, debugging.
- AI/LLM integration — integrating model APIs (OpenAI, Claude), retrieval (RAG), prompt and agent design, and the cost/latency realities of running models.
- Cloud + production basics — Docker, one cloud (AWS is safest), CI/CD, monitoring, secrets, and how systems fail.
- Customer communication — explaining technical tradeoffs simply, managing expectations, writing clear docs.
The roadmap by starting point
If you're a software engineer
You already have the hardest pillar. Focus on:
- Building 1–2 LLM applications end to end and deploying them
- Volunteering for any customer-facing or cross-team work at your current job
- Learning to write and present, not just code
Timeline: 3–6 months of deliberate effort, then start applying.
If you're a solutions/sales engineer
You have the customer pillar. Focus on:
- Deepening real engineering — ship production code, not just demos/POCs
- Owning a project end to end (spec → production), not just the pre-sale slice
- Learning cloud and debugging fundamentals
Timeline: 6–9 months to close the technical gap.
If you're a new grad
You'll need to compress both pillars. Focus on:
- A strong portfolio of deployed full-stack + LLM projects
- An internship or first role with customer exposure
- Targeting new-grad FDE tracks (Palantir and others run them)
Timeline: 12–18 months including a first job.
The 90-day skill plan
Month 1 — Build something real. Ship a full-stack LLM app (e.g., a document Q&A tool or a support assistant) with a real backend, deployed to production, with actual users. This single project demonstrates three of the four pillars.
Month 2 — Make it production-grade. Add monitoring, handle rate limits and failures, optimize cost and latency, write a runbook. Learn Docker and deploy on a cloud. This is what separates FDEs from hobbyists.
Month 3 — Practice the customer side. Write up your project as a case study. Record yourself explaining the architecture to a non-technical person. Take on customer-facing work wherever you can find it. Start preparing for FDE interviews (see our FDE interview questions guide).
How to land the first role
- Target the right companies. AI infrastructure companies (OpenAI, Anthropic, Palantir, Databricks, Scale AI, Cohere) hire the most FDEs. Browse current openings on our jobs board and company profiles.
- Position your experience around impact. Resumes that win lead with "shipped X for customer, which produced Y outcome," not a list of technologies. See our FDE resume guide.
- Use the interview to show ownership. Interviewers probe whether you can own a problem end to end and stay composed with customers. Tell stories where you did exactly that.
The honest truth
You don't become an FDE by collecting certifications. You become one by repeatedly shipping real things for real users and learning to communicate while you do it. Build, deploy, talk to customers, repeat. Do that for a few months and you'll be a stronger candidate than most people with the title.
Start with one project this week. That's the whole secret.