Path 2 of 10, by 2026 demand
Forward Deployed / Applied AI Engineer and Architect
Customer-facing engineers who scope, build and ship LLM solutions inside customer environments, and the architects who guide them.
Demand The fastest-growing customer-facing AI title (third-party, M1, M7). On 2026-11-19 OpenAI listed 40 Forward Deployed Engineering and 128 Applied AI roles; Anthropic 37 Applied AI Architect, 12 Applied AI Engineer and 8 FDE roles (our count, §2.3).
What the role is
Forward Deployed Engineers (FDEs) work inside a customer's systems to deliver production AI applications. OpenAI's post asks the FDE to "own the technical solution end to end, from customer discovery and workflow scoping through architecture, hands-on implementation, evaluation, production deployment, adoption, and handoff" (official, O6). Anthropic's FDEs deliver "MCP servers, sub-agents and agent skills" into production workflows (official, A8).
Applied AI Engineers and Applied AI Architects sit closer to pre-sales and advisory: technical discovery, architecture guidance, eval frameworks, proofs of concept and executive communication (official, A6, A7). This research added the Architect title to the path: it is the largest customer-facing AI group at Anthropic, and its asks (5+ years as a Solutions Architect, enterprise integration, "comfortable with python") match your background most closely.
Level signals
- Senior: delivers a customer deployment with an owner on the customer side and measured adoption.
- Staff: owns the technical win in complex enterprise deals; turns repeated deployments into reusable patterns and product feedback; trusted by customer executives.
- Principal / Manager: shapes the deployment practice: reference architectures, playbooks, hiring bar, partnership with product.
The typical loop
Neither lab publishes an FDE-specific loop. OpenAI's single guide applies to all roles (official, O1, O2). Expect a customer case alongside engineering rounds.
| Stage | What it tests | Basis |
|---|---|---|
| Recruiter and hiring manager | Customer experience, motivation, travel tolerance (up to 50% in one OpenAI post) | Official job post (O6) |
| Skills assessment | Pair coding, take-home or technical test, team-dependent | Official OpenAI guide (O1) |
| Customer case / role-play | Discovery, scoping, objection handling | Common industry pattern; no single source |
| LLM system design | Architecture inside customer constraints: identity, data, residency | Common industry pattern; no single source |
| Coding | Practical Python; sometimes building with a model API | Common industry pattern; no single source |
| Behavioural and values | Ownership, communication, judgement | Official OpenAI guide (O4); third-party for Anthropic (A9) |
Track order
Work top to bottom. Ticks come from the lessons you mark complete.
- 01Working with customersCustomer discovery and technical consulting track. The half of the job most engineers have not practised.
- 02Building blocksLLM and agent engineering track. What you will build in the field.
- 03Production AI systemsLLM and agent engineering track. Evals and guardrails are how you win trust and launches.
- 04Securing AI systemsSecurity engineering for AI systems track. Security reviews gate every enterprise deployment.
- 05The design methodSystem design track. Architecture reasoning for customer constraints.
- 06Worked promptsSystem design track. The enterprise Q&A and gateway prompts are FDE-shaped.
- 07Arrays and stringsCoding patterns track. Keep Python sharp for the coding round.
- 08Behavioural rounds at Staff levelLeadership and behavioral track. Customer-impact and influence stories.
- 09Mock roundsMock rounds track. Do the customer case mock at least twice.
Role practice set
Prompts written for this site. They are not reported interview questions.
- Run the customer case mock (insurer claims) with a partner, then write the one-page POC charter.
- Build a working prototype in an afternoon: an agent that answers questions over a CSV and a PDF, with three eval cases and a README a customer engineer could follow.
- Draw the data-flow diagram for deploying an assistant inside a customer's Azure tenancy with private networking and managed identity; annotate each security question you expect.
- Present a business case to a "CFO" partner in five minutes: cost, value, risk, time to value.
- Write the post-deployment handoff document: runbook, eval suite, owner, escalation path.
Your résumé and story angle
- Enterprise Solutions Architect is the headline. Anthropic's Architect post asks for "5+ years of experience in technical customer-facing roles such as Solutions Architect" and "experience designing scalable cloud architectures and integrating with enterprise systems" (official, A7). Lead with [customer], [problem], [architecture], [outcome].
- Show you still build. FDE posts want "deeply hands-on" engineers (official, O6). Pair each architecture story with code you wrote: [AI integration project, repo, numbers].
- Name the reviews you passed. Security, architecture and procurement reviews are explicitly in the job (official, A6). [A deployment you got through a security review, and how.]
- Azure plus Kubernetes is how many enterprises will consume models in their own tenancy; frame it as removing deployment blockers.
Six-week plan
Tick weeks off as you go
Research notes
Title drift: "Forward Deployed Engineer", "Applied AI Engineer", "Applied AI Architect", "Solutions Architect, AI" and "Deployment Lead" overlap heavily. Read the responsibilities, not the title, and apply across them.
Sources
- Official OpenAI Forward Deployed Engineer (Healthcare)
- Official OpenAI Applied AI Engineer
- Official Anthropic Forward Deployed Engineer
- Official Anthropic Applied AI Engineer, Enterprise
- Official Anthropic Applied AI Architect, Enterprise Tech
- Official OpenAI interview guide
- Third-party JobsByCulture: FDE roles snapshot (2026-07-14)
- Third-party LinkedIn: Jobs on the Rise 2026 press release (2026-01-14)
- Reference Full research notes: docs/RESEARCH.md