A frontier AI lab

Forward Deployment Engineer

Deploy frontier models into the messy real world of the institutions that matter. Then find out where you stand against everyone else who applied.

See where you rankRead the role
4 CVs have been scored for this role
01Drop your CVIt gets read and scored against this exact role in seconds.
02Pick your judgeSix hiring managers, six completely different ways of reading you.
03ClimbVerify your email, interview, and move up the board.
What this job actually is

A Forward Deployed Engineer (FDE) is an engineer who works inside the customer's building instead of behind a product roadmap. Palantir invented the model, and every serious AI lab now runs a version of it.

You are not doing support and you are not doing sales engineering. You own a real business problem end to end: you find it, you scope it, you build the system, and you stay until it actually works in production.

The job exists because frontier models are general and customer problems are specific. Somebody has to close that gap with real data, real compliance, and real people who are sceptical of AI. In 2026 that gap is the commercial bottleneck, which is why demand exploded.

Companies hiring FDEs137
Open roles tracked249
Largest single employerGoogle, 35 roles
Typical experience5 to 7 years
Also posted under these titles
Forward Deployed EngineerForward Deployed AI EngineerApplied AI EngineerDeployment Solutions EngineerForward Deployed Software EngineerSolutions Engineer
Who hires this profile

Companies actively running this role in 2026, grouped by the kind of work you would actually be doing.

Frontier AI labsThey define the category, the pay bands and the way the role is designed.
OpenAIGlobal FDE org deploying frontier models into enterprise and government.
AnthropicDedicated forward deployed engineering team for major customers.
PalantirInvented the FDE model. Still the reference point for the whole market.
MistralEuropean lab building out enterprise deployment squads.
CohereEnterprise-first lab with heavy customer-embedded engineering.
Scale AIOne of the largest FDE employers outside Palantir.
Data and cloud platformsHighest raw volume of open roles right now, often under a different title.
GoogleLeads live job indexes by open FDE role count.
DatabricksLarge field engineering org central to enterprise AI delivery.
Amazon Web ServicesDeployment engineers embedded with strategic accounts.
MicrosoftCustomer-embedded AI engineering across enterprise.
SalesforceApplied AI and deployment roles inside customer teams.
VercelSmaller, product-led version of the same embedded model.
Consulting and defenceWhere the model spread once enterprises copied it. Often more structured delivery.
AndurilDefence tech built largely by Palantir alumni.
DeloittePalantir-adjacent deployment practices at scale.
McKinseyQuantumBlack style embedded AI delivery.
CapgeminiEnterprise AI deployment across industrials.
Could you actually get this job?

Be honest with yourself against these. The more you can say yes to, the more seriously a hiring manager will read you.

✓You have shipped something into production that real people depended on, and you can explain exactly what broke.
✓You have sat in front of a customer or senior stakeholder and changed their mind, not just delivered a spec.
✓You are comfortable when the problem is undefined and nobody hands you requirements.
✓You have touched LLMs seriously: retrieval, agents, evaluation, not just calling an API once.
✓You can travel, and you want the commercial pressure of owning an outcome rather than a ticket.

Usually 5 to 7 years of experience, heavy travel of 25 to 50 percent, and pay above standard engineering bands because the role carries revenue. Note that the market has fragmented: searching only the exact phrase Forward Deployed Engineer will hide most of the openings.

The role

Forward Deployment Engineers are the tip of the spear. You embed with customers, governments, banks, defense primes, hospitals, industrial giants, and turn a raw frontier model into a system that changes how they operate. Half engineer, half consultant, fully owner. You fly in, understand the problem no one has solved, and ship the thing that solves it.

What you'll do
/Map real workflows inside the customer's world and find where a model creates step-change value, a deployment, not a demo.
/Build end to end: pipelines, retrieval, agents, evals, guardrails, UI, whatever gets to production.
/Own outcomes, not tickets. If the pilot fails, it is yours. If it 10x's the customer, that is yours too.
/Translate between a room of researchers and a room of operators, and be trusted by both.
/Come back from the field and make the core product better for the next ten customers.
What we look for
/Strong software engineering: Python or TypeScript, distributed systems, cloud.
/Real experience wiring up LLMs: retrieval, tool use, agents, evals, fine-tuning where it earns its keep.
/Extreme ownership. You run toward the least defined problem in the room.
/You win the trust of a sceptical senior stakeholder in one meeting. And you travel.

This is not a 9 to 5. It is the shortest path from frontier research to real world impact.

The ranking engine

See where you rank

Drop your CV, pick the person who judges it, and get placed on the board. Then earn your way up in a live interview.

01Drop your CVPDF or Word, scored in seconds
02Choose your judgeEach one reads you completely differently

Each judge is a set of written instructions that imitates how a public figure has said they hire. It is not the person, is not affiliated with them, and is not an opinion they have actually given. Only a profile marked Claimed has been confirmed by the real person.

Pick a hiring manager to continue.
Join in

Be the one who judges, or post your own job

Anyone can become a hiring manager here. You write down how you judge a CV, we turn it into your profile, and candidates can choose to be ranked by you.

We send your confirmation link here.

max 40
0 / 40
max 50
0 / 50
max 80
0 / 80
max 700

What earns points, what loses them, who you reject. Be opinionated: this is what makes your ranking different from everyone else's.

0 / 700
max 300

Used when you explain a score or interview someone, so it sounds like you.

0 / 300

Optional

Optional

We review it by hand, publish it, then email you a link to confirm it is really you. Until you confirm, your profile shows as unclaimed.