Prompt Engineering Used to Be a Job Title. Now It’s Everyone’s Job
By Merin Susan John ·
A Java developer who learned the language in 2010 could still ship production code with it today. The syntax has evolved, but the core knowledge has retained its value for over a decade. That’s not how it works anymore, at least not if the skill one has learned has AI attached to it.
“Even your Anthropic certification is valid for six months,” says Srinivas Reddy, Managing Director of EPAM India. “It’s not like you can get certified and do that for the next 10 years. People need to get into that culture, invest time and make it a constant learning thing.”
In India’s tech industry, the credentials people once treated as a finish line are now just a checkpoint. The fundamentals of software engineering—logic, architecture and problem-solving—still hold. But the AI layer on top of them is ageing in dog years, and recruiters have noticed.
Increasingly, they’re hiring for the ability to learn rather than for what a resume already says someone knows.
The Numbers Behind the Anxiety
The World Economic Forum’s Future of Jobs Report 2025 projects that 39% of workers’ core skills will change by 2030, with AI and big data ranked among the fastest-moving capabilities.
Neeti Sharma, CEO of TeamLease Digital, watches this play out in real hiring data. “The fundamentals of AI such as programming, mathematics and problem-solving still stay relevant for years,” she says. “But AI tools, frameworks and platforms are changing so quickly that many of them need to be relearned every six months or so.”
The numbers back that up in an unexpected way: only about 31% of TeamLease’s new hires land in core AI roles. The other 69% go into AI-adjacent positions, jobs that didn’t require AI expertise a few years ago but now do.
“That tells us employers are looking less at certificates and more at people who can learn quickly and adapt as technology evolves,” Sharma says.
Prompt Engineering’s Rapid Rise and Fall
Nothing captures the speed of this churn better than prompt engineering. A year ago, it was the most talked-about title in AI hiring. Companies were recruiting dedicated specialists just to communicate effectively with LLMs.
“The market has moved very quickly,” Sharma says. “Prompt engineering is no longer seen as a standalone specialist role; it's becoming a basic skill that many professionals are expected to have.”
The demand didn't disappear; it simply moved downstream. Companies now want engineers who can take an AI solution beyond the demo stage and into production. “Today, companies are hiring people who can build and deploy AI solutions in production,” Sharma explains. “Skills in agentic AI, MLOps, AI engineering and production deployment are seeing much stronger demand.”
EPAM experienced this transition from the inside. Reddy says the company initially trained employees exclusively in prompt engineering before realising that alone wasn’t enough to stay competitive. It rebuilt its entire AI learning framework around three tiers—AI Ready, AI Practitioner, and AI Champion—each demanding real production experience, not just completed coursework. The scale is telling: over 8,000 AI Ready professionals, more than 4,000 AI Practitioners, and around 450 AI Champions across EPAM’s India operations, alongside more than 1,000 Anthropic Architect certifications.
Not Everyone Buys the ‘Shrinking Shelf Life’ Story
Push back a little, though, and the narrative gets more interesting. At R Systems, the story isn’t that AI skills expire—it’s that they graduate into the basics everyone is expected to know.
“I don’t think there is a shelf life,” says Shardul Sangal, Co-Chief Delivery Officer at R Systems. “It is about learnability. Certain things will become part of your basic curriculum.”
He points to prompt engineering as proof. What was a specialised skill just months ago is now something every junior developer picks up on day one. “Prompt engineering is now part of any developer’s or QA’s job,” he says. “It used to be a skill. Now, if they’re given Cursor AI on day one, they start writing prompts on day one.”
Sangal’s read on where this leads is almost the opposite of “your skills will expire”—it’s “everyone’s skills will merge”. He expects a future where nearly every engineer is, by default, an AI engineer.
“Almost everyone will be an AI engineer,” he says. “It won’t be an exclusivity beyond a point. What will remain exclusive is how much business knowledge you have and how you're able to solve business problems using AI.”
What Actually Survives the Churn
That last point may be the real headline. As AI capability becomes table stakes, what separates one engineer from another is no longer which framework or model they know. It becomes how well they understand the business problem the AI is supposed to solve—and how quickly they can retool when the ground shifts again.
For people just entering the workforce, that changes the whole playbook.
“We expect nearly 39% of today’s core workplace skills to change by 2030,” Sharma says. “The best approach is to build strong fundamentals and keep updating practical skills through real projects, open-source contributions and short certification courses, rather than waiting for formal training every few years.”
So maybe the six-month expiry date isn’t really the point. Certifications will keep losing relevance. Frameworks will keep getting replaced. Models will keep improving on a schedule no one can plan around. What’s becoming durable isn’t mastery of any single AI tool—it’s the discipline of learning the next one faster than everyone else does.