Forward-Deployed Engineer (FDE)
Is this role for you?
Build and own production AI deployments inside a customer’s organization.
- Best fit
- You want to understand a customer’s problem and build the solution with them.
- Poor fit
- You want uninterrupted coding with a stable, predefined scope.
- Coding depth
- Deep: production code, integrations, debugging, and deployment.
- Customer contact
- High: work closely with customer teams; travel varies.
- Experience to expect
- Experienced engineer; OpenAI’s UAE example asks for 5+ years.
- Interview preparation
- Prepare customer problem-solving, integration design, and coding.
Recommended Few-Shot Academy path
I would work through Foundations → Intermediate → Advanced and build the capstones. Then practice scoping a customer project. Add cloud and integration experience separately.
Three differences that matter
- Own the customer deployment through production. [1]
- Integrate AI into existing enterprise systems. [2]
- Turn field lessons into reusable tools and product feedback. [3]
Evidence and limits
I would judge fit by the work described here. The linked postings give specific examples. Check the requirements and interview format for the role you are considering.
Use these interview topics to plan your practice. OpenAI’s posting does not specify how it scores each topic. The graduate routes discussed below have different requirements.
What the job involves
An FDE works with a customer’s team and takes responsibility for getting software running in their environment. That includes writing production code, integrating existing systems, and resolving deployment problems. The distinction to look for in a posting is ownership: will you deliver and maintain the system, or support a sale?
Employers use “forward-deployed” for different responsibilities. Read what you would own before treating the title as a description of the work.
Background: Forward Deployed Engineer on Wikipedia.
A day in the life
The percentages below illustrate a possible split; they are not a measured average. A working demo still leaves integration work: single sign-on (SSO), existing data pipelines, regulatory constraints, and approval to access production systems.
- Customer-facing: 60%
- Requirements analysis with end users, live discovery, iterating in place as edge cases surface.
- Deployment-specific coding: 30%
- Integrating with the client’s existing systems, configuring data pipelines, shipping into a live environment.
- Internal: 10%
- Sharing what customers need with the core product team.
Plan for the work after launch. Watch for changes in model behavior and broken integrations. Turn what you learn into a checklist, template, or internal tool that the next engagement can reuse. Customer requests can also inform the core product roadmap.
The postings cited in this research describe that work directly. Zendesk calls it “bridging the gap” between the field and the product roadmap. Notion asks for “reusable technical assets” and a “feedback loop” into planning. Deloitte expects automated build and deployment workflows (CI/CD), logging, and documentation for the team that operates the system.
Customer scoping, implementation, and ongoing operation all need preparation. When reviewing a role, check which of these responsibilities the employer expects you to have handled already.
Experience and entry paths
Many FDE postings ask for prior engineering or deployment experience. The examples collected for this guide have different requirements:
| Company | Experience floor |
|---|---|
| Salesforce | 3+ yrs (6–10 for Senior) |
| OpenAI | 5+ yrs (8+ for Manager) |
| Anthropic | 4+ yrs |
| Scale AI | 3+ yrs post-grad |
| Databricks | 6+ yrs (15+ yrs big-data exp. at Staff level) |
Other routes and requirements to check:
New Grad tracks
Scale AI and Palantir have dedicated New Grad FDE tracks with different experience requirements.
Founder background
Anthropic also encourages former technical founders to apply.
Federal / public sector
Government-track variants at Palantir, OpenAI, Anthropic, Databricks, and Scale AI can add citizenship and security-clearance requirements. Check each posting for both clearance and experience requirements. The cited Scale AI public-sector posting lists $138K–$219K base.
An early-stage startup engineer may already have relevant experience: talking to customers, choosing a scope, and shipping a working system. Other backgrounds discussed in the research include software engineering, cloud operations, data engineering, machine learning, consulting, and cloud professional services. I would use examples of customer delivery to explain what transfers from your previous work.
Use the skills below to identify gaps in the work you can already demonstrate.
Skills to practice
These six areas organize the technical preparation described in the postings.
Languages: Practice Python and any additional language named in the posting, such as TypeScript, Go, Java, or Salesforce’s Apex.
Data: SQL window functions, common table expressions (CTEs), and query optimization. The research also names Spark, Airflow or dbt, and a data warehouse such as Snowflake, BigQuery, or Databricks. The cited Databricks Staff-level example asks for Scala alongside Python and 15+ years of general big-data experience.
Cloud and infrastructure: prepare the platform named in the posting, such as AWS, GCP, or Azure. Practice Docker, Kubernetes, Terraform, and the cloud networking, access controls, and secret management needed for a client’s security review.
Integration: REST, GraphQL, and streaming APIs; identity protocols such as OAuth, SAML, and SCIM; rate limits and retries. Practice connecting to the customer’s existing sales or business-management systems.
AI and machine learning: applications using large language models (LLMs), retrieval-augmented generation (RAG), which supplies source material to a model, prompt design, and evaluations of system behavior. Prepare to explain agent coordination as well as any framework you use, such as LangGraph, LangChain, CrewAI, or DSPy. The cited Anthropic posting asks for “production experience with LLMs including advanced prompt engineering, agent development, evaluation frameworks, and deployment at scale.”
Testing and operation: prepare automated build and deployment workflows, structured logs, and traces of agent behavior. Version prompts and retrieval indexes so you can reverse a bad change. Explain how monitoring would reveal a problem after launch, and who would act on it.
I would practice customer scoping alongside the technical work. Turn a vague request into a specification, explain what you can deliver, and resolve conflicting requirements from IT, security, and the business sponsor.
The research examples emphasize different skills: data engineering and models of business entities at Palantir, data platforms and full-stack delivery at Databricks, and evaluations, retrieval, agent loops, and prompt versioning at OpenAI and Anthropic. Use the specific posting to choose where to spend your preparation time.
The curriculum covers the AI application layer. Use the gaps below to plan the additional engineering and customer practice you need.
Curriculum mapping
What the curriculum covers
Foundations
The basics of AI, language models, retrieval, and agents needed to understand a deployment.
Intermediate
Retrieval, tool use, and evaluation for a custom enterprise deployment.
Advanced
Build an agent with safety checks, tracing, and evaluations, then prepare it for deployment.
What this curriculum doesn’t cover
- Cloud infra hands-on depth: Docker, Kubernetes, Terraform, VPC/IAM at production scale (this curriculum runs locally via Ollama, not against real cloud infra)
- Enterprise data engineering at scale: Spark, Airflow/dbt, Snowflake, BigQuery
- Enterprise integration protocols: OAuth/SAML/SCIM, legacy ETL, on-prem systems
- General production software-engineering fundamentals most postings assume as a floor (this curriculum assumes zero prior CS background by design: it’s the AI-specific layer, not a CS-degree substitute)
- Customer judgment: breaking down case studies, discovery conversations, and stakeholder management. These need practice beyond the chapters.
Building the projects gives you examples to discuss. Practice explaining the scope and design decisions aloud before an interview.
The interview
Prepare technical work, customer judgment, and reasoning through an unclear problem. The rounds below describe preparation areas from the research; ask your recruiter which formats apply to the role you are interviewing for.
- 1
Recruiter screen
Standard fit and background conversation.
- 2
Coding / technical round
Prepare integration design, debugging, and production-quality code. Confirm whether your employer also tests algorithms.
What to expect, how to prepare, and two things to have ready →
- 3
Case-study round
A hypothetical customer hands you a vague problem and you decompose it live, 45–60 minutes.
Practice scoping before proposing a solution - 4
Behavioral round
Client ownership, accountability under repeat failure, communicating technically to non-technical stakeholders, navigating internal client politics.
- 5
System design (some companies)
At AI-native companies, sometimes a fifth round on agentic/ML system design.
Reported timelines vary: roughly 3–5 weeks at OpenAI, 6–8 weeks at Google, and 3–6 weeks in broader candidate accounts. Confirm the schedule with your recruiter. I would reserve practice time for case studies and communication as well as coding.
When searching for openings, compare the responsibilities behind each title.
FDE vs. the adjacent titles
Is this really the role you want?
Similar titles can describe different work. Use scope and code ownership to compare a posting with the role you want:
| Role | Scope | Code ownership |
|---|---|---|
| Sales Engineer | Pre-sale, ends at contract signature | Sales-first, no production code owned |
| Solutions Engineer / Architect | Pre-sale through early implementation, designs the blueprint | Codes mostly for proof-of-concepts |
| Forward-Deployed Engineer | Post-sale through full production deployment | Ships and owns production code |
Compare compensation at the same employer and level. Check the delivery responsibilities and whether each figure includes a bonus, commission, or equity.
Finding openings
Search the titles below, then compare the responsibilities with the table above. Check whether you would own production delivery, support a sale, or do both.
Places to check for openings:
Direct career pages
Palantir, OpenAI, Anthropic, Google Cloud, Databricks, Scale AI, Salesforce, C3 AI.
Dedicated FDE orgs
OpenAI’s "The Deployment Company" (majority-owned joint venture, $4B+ raised, announced May 2026), AWS’s $1B FDE org (announced June 2026), Microsoft’s "Frontier Company" (roughly 6,000 embedded experts).
Beyond the famous names
Other employers found in the research include Sarvam AI, Talan, Machinify, 3Pillar, and Next League. Check their current openings.
Job boards
startup.jobs (filtered to Forward Deployed Engineer), Built In, and title search on Glassdoor/Indeed.
Recruiting agencies
Perspective AI, KORE1, Paraform, and Recruiting From Scratch publish FDE recruiting material. These are search leads, not recommendations.
Choose certifications based on the platforms named in the posting. The research includes AWS Certified Solutions Architect, AWS AI Practitioner, and Azure AI Engineer Associate or AZ-104/AZ-305 as cloud preparation options. Check the provider’s current exam details before registering.
Platform-specific credentials may be useful when you will deliver on that platform. The cited Salesforce postings name Administrator, Platform Developer I, and Agentforce Specialist as nice-to-haves. Databricks also offers platform certifications.
I would lead with a shipped project and a customer-delivery example. Use certifications to support those examples and address requirements in the posting.
Compare compensation by employer, level, and whether the figure includes equity.
Context, not a headline number
Compensation
| Source | Range |
|---|---|
| General market average (ZipRecruiter / Glassdoor) | ~$116K–$198K base |
| Posting-level analysis (Recruiting From Scratch, n=924) | Median $183K, 25th–75th pct $160K–$215K |
| Frontier-lab total comp (Palantir, OpenAI, Anthropic) | $215K median at Palantir up to $600K–$1.2M+ for Staff/Principal, equity 55–70% of comp at the top |
| Salesforce (published range, JR349466) | $88,970–$287,910 base ($97,860–$316,750 in top metros) |
| Palantir FDSE (estimate) | ~$135K–$200K base + RSUs / sign-on |
| Scale AI, public sector (Honolulu posting) | $138K–$219K |
Compare offers at similar locations and levels. Separate base pay from equity and bonuses, and check the employer’s current posting before using these figures to plan an application.
Comments
Comments are provided by Giscus and GitHub. Loading them connects your browser to GitHub, and GitHub's privacy terms apply.