We are in the midst of a battlefield of AI narratives: consecutive product releases of AI models, media exuberance for the beginning of the agentic AI era, summits discussing AI futures and development challenges, and a relentless arms race between tech companies and AI startups.

Throughout this lively 'AI summer' period of abundant investment and lively media coverage, limited consideration has been given to the expert human labour responsible for developing these state-of-the-art AI systems, products, and services. Thus far, digital labour studies and debates and critical research has extensively looked into gig work, platform labour, microwork. All these different precarious, vulnerable, piecemeal and exploitative forms of work are mediated through digital platforms across the Global South, as well as having significantly increased in recent years across all western societies. 

Hence, there is a sociological blindspot, as Dorschel (2022) describes, when extending the lens of research beyond these precarious types of work. Who are the different expert professionals collaborating to build and maintain AI/ML systems? What are their professional roles, work reality, and inputs in these systems? Recently, a revamped interest in these upper echelons of tech workers is being recorded. Critical scholarship identifies a potential of proletarianisation of data scientists and AI workers following the automation of Machine Learning pipelines (Steinhoff, 2021). Other scholars identify data scientists as a new, hybrid profession consisting of middle class wealth and moral ethics (Dorschel, 2021; 2022). More recent research, borrowing a term of the sociology of culture, perceive data professionals as ‘omnivores' who continuously acquire new skills to remain professionally dominant (Avnoon, 2023; 2024).

Extending this strand of research beyond data scientists, this paper turns the spotlight to different AI & data professionals such as AI and ML engineers, data scientists, data engineers, and other professional roles. In doing so, it hopes to contribute to strengthening our understanding around the professions and professional roles involved in contemporary AI development, and examine the realities and challenges of these novel technical professions. 

The contribution conducts twenty-five semi-structured interviews with professionals working on AI development in the tech industry, start-ups, and academic centres in Ireland. The contribution analyses thematically the discussions with these expert knowledge workers and seeks to provide insights on their professions, expertise construction pathways, the levels of agency and professional autonomy they enjoy, as well as their reflections on generative AI's impacts on their work and the future of their profession.