Is “Forward-Deployed Engineer” Becoming the Default AI Job?
Jason•1d agoI keep seeing some version of Forward-Deployed Engineer everywhere lately.
What’s interesting is that these jobs don’t all look the same.
One role might be a Forward-Deployed Software Engineer building production systems around LLM APIs, agents, and RAG.
Another is basically a Forward-Deployed Full-Stack Engineer expected to ship an entire customer-facing AI product end to end.
Then you have Forward-Deployed Solution Engineers, where the job starts looking more like a hybrid of engineer, solutions architect, and consultant.
And on the more technical side, there are still Applied AI / Research Engineer roles focused on post-training, inference systems, SFT, RLHF/DPO/GRPO, etc.
It makes me wonder whether the AI job market is slowly splitting into two very different tracks:
1. Build better models
Research, post-training, evals, inference optimization, model architecture.
2. Make existing models actually useful
Agents, RAG, tool use, integrations, data pipelines, backend systems, product engineering, and working directly with customers.
And the second category seems to be showing up under the “forward-deployed” label more and more.
The interesting part is that this could actually be a pretty big opportunity for traditional software engineers.
You probably don’t need to become an ML researcher.
But being good at backend / distributed systems may no longer be enough either.
A strong FDE seems to need some combination of:
- backend and system design
- fast product / full-stack execution
- LLM APIs and model behavior
- RAG and retrieval systems
- agent/tool-calling workflows
- evals and observability
- customer-facing communication
- turning vague business problems into something that can actually ship
That raises a few career questions I’m really curious about.
For backend engineers: how difficult is the transition actually? Is the biggest gap AI knowledge, product sense, or being comfortable working directly with customers?
For people already doing FDE work: how much of your job is really AI engineering versus integration / consulting / normal software engineering?
For career progression: where does an FDE go after 3–5 years? Staff engineer? Product? Solutions leadership? Engineering management? Or does FDE become its own long-term technical ladder?
And perhaps most importantly:
How are companies interviewing for these roles?
Are FDE interviews still mostly coding + system design?
Or are we moving toward a completely different loop involving AI application design, debugging an agent, customer scenarios, product cases, and system integration?
Would especially love to hear from people who have interviewed for FDE / Applied AI roles recently.
What company was it, what did the interview loop look like, and what skills did they actually care about?
Discussion
2 commentsSign in to join the discussion
Ask a follow-up question or share practical context.
The career path is the part I’m most curious about. FDE sounds amazing when you’re learning fast, touching lots of customers, and shipping constantly. But after 3–4 years, do you actually build deep technical ownership, or do you end up being the person who is great at rescuing every important customer deployment? I can see it being an incredible path to founding a startup or becoming a product/eng leader. I’m less sure what the long-term Staff+ IC path looks like.
I’ve been wondering the same thing. A lot of these FDE roles sound less like “AI engineer” and more like “really strong SWE who can survive ambiguity and talk to customers.” The AI part is obviously real, but I suspect the harder skill to teach is taking a messy business problem, figuring out what should actually be built, and shipping something that works without needing 6 months of product specs. That probably explains why some of these companies seem to like strong backend/full-stack people more than candidates with a purely ML background.