TL;DR: A solutions architect designs the technical case that wins a deal; a forward deployed engineer builds the production system after the contract is signed. OpenAI posts its architect at $221,000 to $276,000 base plus commission, and its San Francisco FDE at $185,000 to $300,000 with no commission. A sales engineer sits earliest in the deal, at a US median of $124,900.
OpenAI's Applied AI Architect posting calls the job "CTO of your book of business", then hands implementation to delivery teams "once engaged".
Databricks writes the other half into its forward deployed posting in one line: "FDEs are billable." I read the open postings for both roles at OpenAI, Anthropic, Databricks and AWS on October 5, 2026.
- 1Databricks listed 134 solutions architect postings and 99 forward deployed engineer postings out of 892 in total.
- 2Anthropic's London FDE range, £225,000 to £255,000, tops out above every London architect range it posts.
- 3Anthropic's architect posts ask for travel "occasionally"; its FDE posts ask for 25 to 50%.
- 4AWS files its solutions architects under AWS Global Sales and asks them to use Salesforce.
What is the difference between a forward deployed engineer and a solutions architect?
A solutions architect (SA) is a technical advisor who designs how a product will fit a customer's systems before the customer buys.
A forward deployed engineer (FDE) is a software engineer who builds and runs that system inside the customer's environment after the sale.
The architect owns the design and the technical win. The engineer owns the code in production.
The companies hiring both say so in plain words. Databricks' solutions architect posting opens by asking for "experienced pre-sales professionals" who work alongside the Enterprise Account Executive.
Its public sector FDE posting says the engineer will "own the architecture, lead design decisions, and implement end-to-end systems".
- Partners with an account executive on a named book of accounts
- Runs discovery, demos and proofs of concept
- Draws the reference architecture
- Owns the technical account plan and expansion
- Works inside the customer's systems
- Writes and reviews production-grade code
- Owns the rollout to production
- Measured on production adoption
Our role guides cover each job on its own: the solutions architect guide and the forward deployed engineer guide. The architect guide lists pre-sales as one specialization among several. At AI and data platform companies it is the main one.
Where each role sits in the deal
The contract signature is the handoff. Anthropic's Applied AI Architect posting describes "a Pre-Sales architect" who partners with account executives.
Its London FDE posting puts the engineer next to "our Post-Sales, Product, and Engineering teams".
Anthropic now writes that handoff into its org chart. It is hiring a first Pre-Sales Program Lead inside Forward Deployed Engineering.
That person will "work hand in glove with a Post-Sale Program Lead counterpart who owns delivery from signature onward".
The posting asks for "enough technical fluency to scope and defend AI and software services work" without "being the one who builds it".
Stage 8 is where the lines blur. OpenAI's AI Architect posting keeps the architect "accountable for technical progress and customer outcomes" after the sale. It also says delivery teams "own implementation execution once engaged".
The architect keeps the account plan. The build goes to someone else.
Databricks splits the architect job in two around the same line. Its Delivery Solutions Architect is "the post-sale technical lead", who takes over after the "technical win". So a title with "solutions architect" in it can be post-sales.
The difference from an FDE is that the delivery architect drives adoption of a product already chosen, while the FDE writes the custom system.
Where the sales engineer fits
A sales engineer sits earliest of the three. The US Bureau of Labor Statistics defines sales engineers as workers who "sell business products or services, such as software or support, that require technical expertise".
Most of those jobs are outside software. Merchant wholesalers of durable goods employ 22% of US sales engineers and manufacturing 21%, against 6% at software publishers. The software jobs pay the most:
Inside AI companies the title is fading into the architect one. OpenAI had only three postings with "Solutions Engineer" in the title, all in its ads business.
Anthropic's architect postings list "Solutions Architect, Sales Engineer, or Technical Account Manager" as equal routes in.
Skills: how much code each role writes
Both roles now code, but to different finish lines. Anthropic tells its architects to "ship working code" and "write near-production examples that customers can extend".
Its FDEs "work within customer systems to build production applications". The architect's code proves the idea. The engineer's code runs the business.
The coding bar in the postings follows that split. Databricks asks its architects for "production programming experience in Python, R, Scala or Java". It notes they work with "hands-on resources who will build proofs of concept".
OpenAI's FDE posting asks the engineer to "write and review production-grade code across frontend and backend".
| Posting | Minimum experience | Travel |
|---|---|---|
| Anthropic Applied AI Architect, Commercial | 3+ years | "Occasionally" |
| AWS Solutions Architect, Commercial Midmarket | 2+ years | Not stated |
| Anthropic FDE, London | 4+ years | 25 to 50% |
| OpenAI FDE, San Francisco | 5+ years | Up to 50% |
| Databricks FDE, Public Sector | 5+ years | 20 to 25% |
The architect needs more range in the room. OpenAI wants its architect to build trust with "CTOs, CIOs, CISOs, AI leaders" and to own a technical account plan.
AWS adds a sales system to the toolkit: its Commercial Midmarket SA posting lists "salesforce.com and other internal Amazon systems" among the duties. No FDE posting I read mentions a CRM.
The FDE needs depth in the customer's stack. Databricks' AI FDE posting asks for production GenAI work, "including RAG, multi-agent systems, Text2SQL, fine-tuning".
It says the role is "not intended for internship, new graduate, or entry-level applicants". For the AI-specific version of the job, see our AI forward deployed engineer guide.
Pay: base salary, commission and on-target earnings
At OpenAI the architect's pay carries a sales component, and the FDE's does not. OpenAI lists "Offers Commission" on its AI Architect posting and only equity on its FDE posting.
Anthropic says the ranges on its sales roles are on-target earnings that include "sales commissions/sales bonuses". So an architect's posted range can include commission.
Read the bars with that in mind. OpenAI's architect figure is base, with commission on top.
Anthropic's $240,000 to $315,000 architect range is on-target earnings if Anthropic treats the role as sales, which its posting does not state. Databricks says its ranges are base pay for some roles and on-target earnings for others.
It does not say which applies here. AWS lists $151,000 to $204,300 for San Francisco and $131,300 to $177,600 for Seattle.
London gives the cleanest comparison, because Anthropic posts both roles there with figures:
Outside the US and UK, our forward deployed engineer rate card tracks hourly rates by market. Our FDE cost-to-hire page adds the employer costs on top of salary.
Demand: which role companies post more
Databricks and Anthropic post more architects than forward deployed engineers, while OpenAI posts 24 of each:
Anthropic's ratio stands out: 31 architect titles against 6 FDE titles. Its London FDE posting calls new hires "one of our founding FDEs", so the team is still small next to the pre-sales bench.
OpenAI's FDE titles span Tokyo, Seoul, Singapore, Sydney and Abu Dhabi as well as US cities.
Databricks shows the whole chain, because it posts pre-sales, forward deployed and post-sales roles side by side:
Pre-sales architects lead, forward deployed engineers come second, and post-sales delivery architects third. The FDE count is the newer story.
In June 2025, a16z partner Joe Schmidt described the role as a professional services job "sometimes rebranded as a forward deployed engineer". Databricks' own AI FDE posting says the team delivers "professional services engagements".
Career paths between the two roles
The roles feed each other in both directions. Anthropic's International Policy architect posting accepts four years as a "Customer Engineer, Forward Deployed Engineer, Sales Engineer, Solutions Architect".
Its FDE manager posting asks for 10+ years "in software engineering, solutions architecture, or a technical customer-facing role".
The usual FDE ladder in our role guide runs FDE, senior FDE and principal FDE, then technical solutions architect. That last step trades build time for account scope.
The architect's ladder runs toward enterprise architect or pre-sales management. Anthropic's strategic architect posting lists Field CTO as a feeder role.
Anthropic's architect manager roles ask for 7+ years as a solutions architect or sales engineer.
The bar is higher going from architect to FDE. OpenAI's FDE posting asks for 5+ years of engineering or technical deployment experience, with production code across frontend and backend.
Our deployment strategist vs forward deployed engineer comparison covers the other half of a forward deployed team.
Which role a project needs
Hire the architect when the problem is winning or shaping the deal. Hire the FDE once the customer signs and the product does not yet do what the customer needs.
If the product already fits, a post-sales architect or customer success team can carry adoption without a builder.
Early-stage teams often blend the top row into one job. Anthropic's commercial architects are told to "go deep with priority accounts as a hands-on builder".
Once custom builds pile up after signature, all three AI companies split the work and post the FDE as its own role.
Hire forward deployed engineers and architects from Asia
Second Talent matches you with vetted forward deployed engineers and solutions and software architects in 24 hours, at 50-70% below US cost. Each placement carries a 90-day, one-time replacement guarantee.
We employ them through our own EOR in 9 Asian markets, including Vietnam and the Philippines.
Our monthly plan is $4,999 a month for a senior engineer, all-inclusive. Tell us which role you need and we will send a shortlist.
Frequently Asked Questions
Is a solutions architect a sales job?
Often, yes. AWS places its solutions architects in AWS Global Sales, and OpenAI pays commission on its architect role. Delivery architects at Databricks are the exception: they start after the technical win.
Does either role need a cloud certification?
None of the postings cited here require one. Databricks lists its own certification as "nice to have" for architects, and none of the FDE postings mention a certification.
Cloud certificates still help architects at cloud vendors, as our solutions architect guide notes.
Is a forward deployed engineer a consultant?
Close to one, but with product code. Databricks calls its FDE engagements professional services and its FDEs billable. Unlike most consultants, FDEs feed fixes back to their own product teams.
Our forward deployed consultant guide covers the less technical version.
