Forward deployed engineers from Palantir, OpenAI or Anthropic work for the vendor, and their job is to make that vendor's product succeed in your organization. They are excellent at that, and often included with a large platform or model commitment. An independent forward-deployed team has no product to sell, so it builds on the stack you already run and can choose, or mix, models and clouds on the merits. Use vendor FDEs to roll out a platform you have already bought. Use an independent team when the system is yours to design and own.
Palantir made "forward deployed software engineer" a well-known title. Its careers pages describe the role as engineers working directly with customers on their hardest problems. In its engineering blog, Palantir explains the internal nicknames: product engineers are "Devs," and forward deployed engineers are "Deltas," a leftover from early business-development teams named after letters of the NATO alphabet. A Dev focuses on one capability for many customers; a Delta focuses on one customer and many capabilities. Palantir also has a related role, the deployment strategist (internally "Echo"), which sits closer to product management.
The important point for buyers: Palantir's FDEs deploy Palantir's platforms, such as Foundry and AIP. The configuration, data models and applications they build are built on those platforms.
OpenAI has built a forward deployed engineering team and posts FDE roles in cities including San Francisco, Seattle, New York, Munich and Tokyo. Its postings describe FDEs owning a customer deployment from discovery and technical scoping through system design, build and production rollout, working directly with the customer's engineering and domain teams, and feeding what they learn back into OpenAI's products.
Anthropic hires forward deployed engineers in its Applied AI organization. Its postings describe hands-on deployment work with large customers, building production systems on Claude, including integrations such as MCP servers and agent workflows, and documenting repeatable deployment patterns for Anthropic's own product and engineering teams.
Cloud providers and enterprise software companies have followed, with FDE postings of their own. The pattern is the same: a vendor puts engineers next to its largest customers to turn a purchase into daily use, and to learn what the product should do next.
None of this is a criticism of the people. It follows from who pays them.
These models are not exclusive, and many enterprises use both. A common setup for enterprise AI looks like this:
The independent team is the one accountable for the result working end to end. The vendors are accountable for their products working.
If the answers are clear and you are comfortable with them, vendor FDEs may be all you need. If not, you need an engineering team of your own for that work, either hired or brought in.
Ryz forward deployed engineers are an independent team. We work as AI pod teams of senior engineers, typically around seven people including a tech lead, an ML engineer and backend engineers, on your team, in your cloud accounts and repos, on US business hours including New York hours. Our pods work with Anthropic and OpenAI models on AWS and Azure, with GitHub and Postgres, and they pick tools based on your evaluations rather than a partnership. You own all the code and IP.
That is how our pods shipped marketing compliance review for a global capital management firm, using both Anthropic and OpenAI models on Azure, and fraud detection for a global fleet company that confirmed $5.94M in fraud across 244K+ scored invoices. More in our case studies.
We are not the right fit if you want a proprietary AI platform you license and configure. In that case, buy the platform and use the vendor's FDEs to roll it out. For a wider comparison of providers, see the best forward-deployed engineering firms and partners that take an AI pilot to production. If you are weighing model providers themselves, our guide to OpenAI vs Anthropic for enterprise covers that choice.
A software engineer, internally called a "Delta," who works directly with one Palantir customer to deploy and build on Palantir's platforms. Palantir contrasts it with product engineers ("Devs"), who build one capability for many customers.
According to OpenAI's job postings, they own customer deployments from discovery and scoping through system design, build and production rollout, working with the customer's engineering and domain teams and feeding lessons back into OpenAI's products.
Use the vendor's FDEs to roll out a platform you have already chosen. Use an independent team when you need a system built across several vendors and your own systems, and you want to own the design and the code.
Yes, and it is common. The vendor's engineers handle their product; the independent team owns the end-to-end system, including integrations, data work, evaluation and monitoring.
Whatever you run. Our teams commonly work with Anthropic and OpenAI models on AWS and Azure, with GitHub and Postgres, and choose per task based on your evaluation results.
Questions we didn't answer? Email info@ryzlabs.com.
Thanks — your message has been sent. We’ll get back to you soon.
Something went wrong while sending your message. Please try again or email info@ryzlabs.com.