Sarah Khalid was three years into her career as a Success Architect at Salesforce when her team restructured and she found herself in a role she hadn’t planned on, couldn’t fully define, and wasn’t sure existed yet. “I didn’t plan on becoming an FDE,” she said. “It’s something that I grew into.”
Forward deployed engineers (FDEs) saw an 800% spike in job postings between January and September 2025, according to an analysis by Indeed and the Financial Times, as companies adopt AI agents at breakneck speed but need expertise to actually deploy them. FDEs exist to close that gap. They code, consult, and translate agentic AI into working solutions, often while sitting side by side with the customer. Similar to when Harvard Business Review deemed data scientist “the sexiest job of the 21st century” due to a rare combination of technical chops and business savvy, both roles prove that transformative technology doesn’t eliminate jobs. It creates entirely new ones we couldn’t have imagined a few years prior. Now history is repeating itself, just faster.
How do I know when AI is wrong, when AI is right, how it can solve the right problem? That truly is human-AI collaboration.
Ruth Hickin, VP of Agentic Workforce Strategy and Innovation at Salesforce
Remember when ‘deployment’ meant pushing code live?
For FDEs, it means packing a laptop and going to the client site. The distinction matters more than it sounds.
“Implementation teams build solutions,” said Khalid, who is now an FDE Director at Salesforce. “The FDEs make sure those solutions drive value.” A traditional engineer builds in isolation and hands the product off. FDEs don’t arrive with a checklist of predefined deliverables. They align to a customer’s business outcomes and stay accountable until those outcomes are met.
Ruth Hickin, VP of Agentic Workforce Strategy and Innovation at Salesforce, frames what makes deploying agents different: “This is not a typical implementation of software. It’s really, how do we help people get the value and the outcomes of AI? Not just implement it but understand what problems they’re trying to solve.” The result, as she describes it: “kind of like hacker meets customer value meets professional services person.”
The skills that actually matter
This hybrid mash-up of different roles mirrors Khalid’s path. Her move into the role from developer to architect to FDE wasn’t a straight line so much as a cumulative one. “[My] developer side taught me depth, and the architect side taught me perspective,” she said. “The FDE role combines both.”
Agentforce demanded new skills too, she said. “The agentic platform itself is different. It’s not quite as black-and-white as coding. It’s a little bit more art and not as deterministic as writing code would be.” Prompt engineering, instruction-writing, comfort with nondeterministic outputs were not in her prior toolkit.
Then there was the pace. “I was used to three releases a year,” Khalid said. “But the pace of innovation with Agentforce is so much more: weekly, biweekly. A high ability to adapt isn’t a bonus quality for this role. It’s a baseline requirement.”
Coding, as Khalid puts it, is “table stakes.” What distinguishes strong FDEs are qualities that don’t show up cleanly on a résumé: judgment, pattern recognition, and the ability to deliver hard truths to C-suite stakeholders without alienating them. Not to mention, a learning mindset and the ability to share those learnings with others.
Hickin also frames human judgment as the key differentiator in a world of AI-generated output: “How do I know when AI is wrong, when AI is right, how it can solve the right problem? That truly is human-AI collaboration.”
Curiosity over credentials
Salesforce tripled its FDE team in just six months, built from three internal teams: engineering, professional services, and customer success. That origin shapes what the company is actually looking for.
“Sometimes you look at an FDE in the market and it feels like an engineering job, very, very technical,” Hickin said. “But I do think it’s curiosity over credentials here [at Salesforce]. The most important thing is, can you solve a customer’s problem?” Around 40% to 50% of FDE movement at Salesforce is internal. The company built a six-week onboarding program in September 2025 with two weeks of focused technical training and four weeks on the job to bridge the gap.
For anyone eyeing the role from the outside, Hickin’s advice is simple: “Become the AI expert in the job that you’re in. Experiment with the tool, figure out how it can solve problems in your team. Then you become a kind of quasi-FDE within your own job.”
Trust builder
AI has already changed what the job demands internally. FDEs once spent up to 40% of their time on administrative preparation, including summarizing meetings, pulling account histories, and drafting status updates. That’s largely automated now. “Now they can spend their time on hacking, building, figuring out the customer issue,” Hickin said. “Human jobs become more human.”
Khalid explained what this looks like in practice: She once worked with a tech customer whose AI agent was targeting a 70% deflection rate for their customer service function. That figure became the North Star to add functionality, optimize, and move the needle. When they hit it, the customer expanded to a second use case.
She also described a time when an insurance company demanded accuracy and response time targets. Khalid admitted that the technical side was hard. It required experimenting with different LLM models and prompt engineering on the back end. But what nearly derailed the engagement had nothing to do with code.
This company, like many others, was experimenting with a variety of solutions. So she said, “It also required a lot of nontechnical skills: how to deal with that customer, how to make them not want to give up on Agentforce. And make them trust.” That account didn’t turn on a technical breakthrough. It turned on rebuilding confidence. “There’s not much coding involved,” Khalid reflected. “It’s actually more judgment, trust, and communication skills. And not something that AI can replace.”
Work, redefined
Back in 2012, the data scientist role looked like something only a narrow slice of specialists could do. A decade later, data literacy had become a baseline expectation across job functions. Hickin sees the FDE role charting the same arc. It won’t be a specialty forever but a leading indicator of where technical work is heading: closer to the customer, more dependent on judgment, and more human than the job title implies.
Khalid, reflecting on the FDE acronym, has her own proposal: future defining engineers. “We’re not just defining the AI part,” she said. “We’re also defining how this role itself will shape out in the future.”
It’s clever. And after what happened with data science, it’s hard to argue with.
