Dive Brief:
- Skill requirements for AI-focused roles don’t match listed titles, according to an analysis of almost 50,000 job descriptions by tech talent company Andela. Among over 1,800 postings for AI engineer, ML engineer and other similar job titles, more than half conflated skill sets associated with at least two roles, including LLM orchestration, autonomous agent architecture and vector database design.
- The industry is inventing roles faster than it can name them or accurately match skills to job titles, the analysis found. Andela identified several new roles and skill bundles, including “LLM application engineer,” a contemporary AI engineer who builds on foundation models rather than training them.
- “Companies are saying, ‘I need to go hire an AI engineer,’ but that’s an extremely broad term,” Cory Hymel, head of research at Andela, told Channel Dive. “They’re looking for skill sets that do not map to typical back-end engineering roles. And they’re putting the wrong butts in the wrong seats, because no one really knows what these underlying skill sets need to be and what to call them.”
Dive Insight:
That AI is changing the job market is undeniable, but how is often unclear.
“There’s tons of chatter that AI is deleting jobs, AI is changing jobs, but it’s been kind of hand-wavy and anecdotal,” Hymel said. “If these jobs are changing, what jobs are changing? And what are the underlying skill changes that are causing this heartburn in the market?”
Andela’s analysis found that job descriptions for product managers, front- and back-end engineers, developers and other familiar titles are changing and overlapping. Machine learning engineers are being asked to have software architect or back-end engineer skills, for example, and DevOps experts are expected to know about cloud engineering.
Front end engineering is experiencing a major evolution, according to Hymel. “Companies are pushing for front-end engineers to have better soft skills and be able to define and develop features,” he said. “They want front-end engineers to dip a toe into the product management role.”
Hymel called these emerging and overlapping requirements “skill bleeds.” The job description analysis led his team to create several emerging roles, including DevSecOps security engineer, docs-as-code engineer and lakehouse analytics engineer.
Another novel position is FinOps reliability engineer, which is someone who can track cloud infrastructure costs alongside token consumption budgeting and cost forecasting and modeling for AI systems.
Evolving job titles can be a headache for job seekers. Companies previously hired individuals with talents that were useful for the duration of a career, but AI is making some skills obsolete, Hymel said.
“The pace of the technology and the skills required are maturing so quickly that employers are starting to accumulate talent debt in ways they’ve never had to deal with before, and that obviously then gets expressed into job change,” he said.
It’s essential for companies to avoid knee-jerk reactions like mass layoffs, Hymel added. “What we’ve seen out of the data here is that there’s actually a need from enterprises for new areas of work that they’re having a hard time expressing or writing job posts for.”
