FDE Jobs at OpenAI: Roles, Salary, and How to Apply

OpenAI runs the fastest-growing Forward Deployed Engineering organisation in the world — and the best-paying. At the time of writing, our index shows around 30 distinct FDE-titled openings at OpenAI across three continents, pulled directly from their careers feed. If you're targeting one role in the FDE market, this is the one most people target. Here's the complete picture.

The roles

OpenAI's forward-deployed org (largely under Model Deployment for Business) hires across several variants:

• Forward Deployed Engineer (FDE) — the core role: embed with a customer, ship production applications on OpenAI models. Posted in SF, NYC, Seattle, and most international hubs.
• Forward Deployed Software Engineer — closer to full-stack product building inside customer engagements.
• Platform Engineer, Forward Deployed Engineering — builds the internal tooling and infrastructure the FDE org deploys on; less travel, more systems.
• Technical Deployment Lead (FDE) — senior, account-owning; part architect, part engagement lead.
• Forward Deployed Security Engineer / FDE, Gov — specialised tracks for security-sensitive and government deployments.
• Manager, Forward Deployed Engineering — the management fork, posted in SF, NYC, London, and Munich.

Day-to-day, the core FDE job is what we describe in What is a Forward Deployed Engineer? at the AI-native extreme: discovery with executives, building evals before features, RAG/agent architectures against messy enterprise data, and carrying pilots into production under real deadlines.

The pay

OpenAI is the top of the FDE market (full market data here):

Two notes. First, OpenAI publishes salary ranges on its postings — you can verify current numbers on any live listing. Second, equity (PPUs) is a large fraction of the package and its terms differ from standard RSUs; understand the liquidity mechanics before comparing offers.

Locations and travel

US: San Francisco (largest concentration), New York, Seattle. International: London, Dublin, Paris, Munich, Madrid, Stockholm, Tokyo, Seoul, Singapore, Sydney, UAE — the international postings are often the less competitive way in, if you're in-region.

Most roles are hybrid: office-anchored with travel bursts around customer milestones rather than constant road time (how hybrid FDE actually works). Government-track roles carry heavier on-site expectations.

What OpenAI looks for

Across postings and candidate reports, the consistent bar:

1. Production engineering strength. Python and TypeScript dominate. This is a ship-real-systems loop, not a research one.
2. LLM production literacy. Evals as a discipline (golden sets, regression harnesses), RAG architectures, agent/tool-use patterns, structured-output reliability, token-cost engineering. "Built and shipped" beats "experimented with" — make it the top of your resume.
3. Executive-grade communication. FDEs demo to C-suites and defend architectures to customer engineering teams in the same week.
4. Deployment scar tissue. Evidence you've carried something ambiguous into production inside an organisation you didn't control — the exact experience the decomposition round probes.

The interview loop

Expect four stages over roughly 3–5 weeks: recruiter screen → technical screen (practical coding, often LLM-flavoured) → virtual onsite of 4–5 rounds (coding, ambiguous problem decomposition, customer scenario, behavioural) → team matching. Our FDE interview questions guide maps to this loop almost one-to-one; the decomposition and customer-scenario rounds are where most candidates fall.

Preparation that moves the needle specifically for OpenAI:

• Build one end-to-end LLM project with a real eval harness and be ready to defend every design decision.
• Practise decomposition prompts phrased as AI deployments ("a claims processor wants to automate document intake — what do you build first, and how do you know it's safe?").
• Have a crisp answer for "why forward deployed at OpenAI" that isn't about the brand — interviewers screen for people who want the customer edge, not proximity to research.

How to apply

The honest playbook:

1. Apply direct to the specific role/location that fits — current OpenAI FDE openings are here, synced hourly from their feed, with apply links straight into their ATS.
2. Consider the international offices if you're in-region — the London/Munich/Tokyo pipelines are newer and less saturated than SF.
3. A referral helps more than usual. The FDE org is well-networked with Palantir alumni; if you've worked with ex-FDSEs, ask.
4. If you're not ready yet, the classic feeder paths are 1–2 years of FDSE work at Palantir, deployment engineering at Scale or Databricks, or shipping LLM products anywhere with real users — then re-apply. The career-path guide covers the timing.

Compare OpenAI against the rest of the market in Top Companies Hiring FDEs in 2026, or go straight to the live feed.