The breakdown of 2026's hottest AI job title. Who's hiring, what FDEs actually do, the salary, and the 90-day plan to position yourself for one (even if you're not a traditional engineer).
There's a new role exploding in tech right now, and even if you're not technical, this is the moment to position yourself for what's coming next. It's called the Forward Deployed Engineer (FDE).
Salesforce committed to building a team of 1,000 FDEs. Anthropic and OpenAI just launched billion-dollar joint ventures to put engineers inside their customers' companies. Job postings for this role grew over 800% in 9 months last year, and 729% year-over-year by April 2026. Median total comp is $173K. At Palantir it's $215K. At OpenAI and Anthropic, $350K to $550K for mid-to-senior.
This guide is the full breakdown of what an FDE actually does, how someone without a traditional engineering background can position themselves for this role (or its close cousins), the skills that matter, and a 90-day plan to get there.
An FDE goes inside a company, learns how the business actually works, and builds custom AI solutions to fit it. Part engineer, part consultant, part product manager. The role exists because the AI skills gap is the single biggest barrier to enterprise AI adoption (per Deloitte's State of AI 2026 report). The models are ready. The workforce isn't. FDEs close that gap on a per-customer basis.
Day-to-day, an FDE will typically:
1. Run scoping workshops onsite with the customer to map their workflows, data, and constraints.
2. Write agent instructions and iterate on prompts inside the customer's environment.
3. Build and configure data pipelines (Snowflake, Databricks, Foundry-style data models).
4. Ship integrations between systems that weren't designed to talk to each other.
5. Build evaluations and run A/B tests on LLM workflows for the specific use case.
6. Debug production incidents and write the root-cause analysis for both engineers and executives.
7. Feed product feedback back to HQ so the core product gets better.
The job market analyst Bloomberry analyzed 1,000 real FDE job descriptions. "Working directly with customers" appeared in 55% of them. "Build/deploy AI/ML" in 37%. "API integration" in 32%. Sales quotas in 0%. This is not a sales role disguised as engineering.
FDE jobs use different titles across companies. Search for all of these on LinkedIn:
Palantir (Forward Deployed Software Engineer, FDSE)
OpenAI (Forward Deployed Engineer)
Anthropic (Forward Deployed Engineer, Applied AI Engineer)
Salesforce (AI Forward Deployed Engineer, Agentforce FDE, Deployment Strategist)
Databricks (AI Engineer, Customer-Facing Engineer)
Cohere (Forward Deployed Engineer)
Ramp (Forward Deployed Engineer, Senior FDE)
Rippling, Intercom, Scale AI, C3 AI, Box, Latent Labs, Lindy, Commure (FDE)
Deloitte and EY (AI Forward Deployed Engineer, launched their FDE practices in 2026)
Google Cloud (Customer Engineer, Applied AI Engineer)
58% of FDE roles are at growth-stage startups with 11 to 200 employees. New York (35% of postings) has now overtaken San Francisco (11%) as the primary FDE hub.
Salary expectations
Median FDE total comp in the US is $173K (Bloomberry analysis of 1,000 jobs). Palantir FDSE: $171K to $415K, median $215K. Anthropic and OpenAI: $350K to $550K for mid-to-senior. Staff-level clears $630K+. UK roles are £108K to £253K. New-grad TC across the category is $180K to $250K. 70% include equity. Zero include sales quotas.
Let's be honest. The data says 45% of FDEs come from software engineering, 22% from Solutions Engineering, 15% from data engineering or data science. Only about 18% come from non-traditional backgrounds. Top-paying firms expect production-level coding.
But there are two real paths if you're not a traditional engineer.
Path 1: The adjacent-title path. These roles do FDE-equivalent work with a lower coding bar. Search for:
Solutions Engineer (AI), AI Implementation Consultant, Customer Engineer, Deployment Strategist, Applied AI Specialist, Forward Deployed Consultant, AI Integration Lead.
Salesforce's pod model literally pairs a Deployment Strategist (non-coding, business-focused) with two FDEs. Anthropic, OpenAI, and Deloitte all hire "Implementation" and "Applied" roles that emphasize business framing + AI fluency + light coding. Aim here first.
Path 2: The pivot-and-prove path. Build coding fluency to the level where you can ship one real integration, one real RAG or agent app, and one written debug case study. SkillScouter's framing: "FDE hiring is portfolio and performance-driven, not credential-driven." People come from bootcamps, non-CS degrees, and self-taught backgrounds. But you have to actually ship.
Either path is real. Don't sell yourself a fantasy that you'll land a $400K Anthropic FDE seat without writing Python. Do start in the adjacent-title bucket and earn your way up.
1. Python. Production-level, not notebook-level. Appears in 66% of FDE job descriptions.
2. SQL and data modeling. Snowflake, Databricks, BigQuery, dbt patterns. Know how to write queries that actually work on real data.
3. LLM orchestration. RAG (retrieval-augmented generation), evals, agent design, vector stores. RAG appears in 12% of JDs explicitly; AI Agents in 35%; LLM in 31%.
4. One cloud well. AWS (32%), GCP (22%), or Azure (18%). Pick one and go deep. Don't dabble across all three.
5. API integration. With real auth, retries, and error handling. Not "I called an API once." Real production patterns.
6. Prompt engineering and eval-writing. For production LLM systems, not just chatbots. Show you can measure quality, not just demo it.
7. Customer interviewing and requirements elicitation. The ability to scope a problem with a customer in 60 minutes without doing 5 weeks of discovery.
8. Business-problem framing. Translating ambiguous customer pain into a shippable scope. This is the part that separates FDEs from "junior engineers."
9. Executive communication. Explaining a 12-line bug to a CFO. Writing a one-page incident report a board would understand.
10. Building with AI coding agents. Cursor, Claude Code, Codex. Modern FDE leverage is 5x to 10x via these tools. If you're not using them daily, you're behind.
Days 1 to 30
Get Python and SQL to working production level. Take one course (DeepLearning.AI's LLM specialization is a good anchor). Set up Cursor or Claude Code and build with it every day.
Start one integration project: pull data from one real system into another with auth and error handling. Example: pull Stripe customers, enrich them with one HubSpot field, post a Slack alert. Use real APIs, not mocked.
Join FDE Academy's community, subscribe to the Pragmatic Engineer newsletter, and follow Aaron Levie, Sundeep Teki, Gergely Orosz, and Palantir alumni on LinkedIn and X.
Days 31 to 60
Ship a RAG app to real users. Even 10 users counts. The app should answer questions from a real document set (a company knowledge base, your own notes, a public dataset).
Write one eval suite for an LLM workflow. Show you can measure correctness, not just demo it.
Complete your integration project from Days 1 to 30. Publish a public GitHub repo with a clean README, and record a 5-minute Loom walkthrough.
Do 5 mock customer-discovery calls. Record yourself scoping a fake AI project with a friend playing a CFO or CMO. This is what FDE interviews actually test.
Days 61 to 90
Publish a debugging case study. 700 to 1,000 words. "What broke, how I investigated, what I changed." This is the highest-leverage portfolio piece per every FDE coaching resource. Almost no candidate has one.
Rewrite your resume in deployment-outcome language. Not "worked on AI projects." Try: "Deployed X used by Y users; handled Z auth edge case; reduced inference cost by N%."
Apply to 15 to 30 roles. Use the title list above (FDE, Applied AI Engineer, Customer Engineer, Solutions Engineer AI, Deployment Strategist, AI Implementation Consultant).
Post 1 to 2 case-study posts per week on LinkedIn. Hiring managers screen for the "founder/operator" signal. Posting consistently is the cheapest way to build it.
Every credible FDE coaching source converges on these three artifacts. Build all three by Day 90.
1. One integration project. Two systems that weren't designed to talk. Real API, real auth, real error handling. Deployed, not mocked. One real user.
2. One deployed AI pipeline. RAG app over a real document set, or an agent doing a real job, or an LLM workflow processing real data. Production, not notebook. Eval suite included.
3. One written debugging case study. The narrative of "what broke, how I investigated, what I changed." Most differentiating piece you can have. Almost no one writes one.
Where the jobs are posted
LinkedIn Jobs (search the 7 titles above + add the LinkedIn Skills filter for "LLM," "RAG," "Agent Development"). Greenhouse boards for Anthropic, OpenAI, Ramp, Rippling, Cohere, Databricks. careers.salesforce.com for Agentforce FDE roles. Levels.fyi job board (filter by comp). fwddeploy.com (specialist FDE platform). Y Combinator's Work at a Startup for series A-B FDE roles.
The Only AI Masterclass You Need
If this guide helped, but you’re looking to go deeper, I got you!! My 30-Day Challenge takes you from saving AI tips you never use to actually building with AI, step-by-step.
I show you exactly how I automated two e-commerce brands, my social media, and most of my personal life, then hand you the agents, workflows & systems to do the same. I’m teaching you every single thing I know with one lesson and one build a day.
Join the AI Masterclass →© 2026 Mariah Brunner. All rights reserved.