Before starting my new role, I already knew it would be highly cross-functional. But after working closely with Claude as a daily work companion, I’ve found it increasingly difficult to define what role I am actually performing.
Sometimes I am doing full-stack work: thinking through frontend interactions, backend logic, and data flow.
Sometimes I am doing pipeline design: figuring out how an agent’s capability can be reused, composed, and reliably invoked.
Sometimes I am doing data science or data engineering: looking at data, identifying signals, designing evaluation frameworks, and understanding system bottlenecks.
Sometimes I am doing model evaluation: asking where the model is strong, where it is weak, and how much stronger or weaker it is across different scenarios.
Sometimes I am doing technical writing: writing skills, documentation, and turning tacit knowledge into reusable workflows that others can build upon.
And sometimes, unexpectedly, I am also doing product thinking, art direction, and creative strategy.
In the past, job boundaries felt relatively stable. Frontend engineers built the interface. Backend engineers managed the logic. Data scientists worked on models and analysis. Product managers defined requirements. Designers shaped the experience.
But now, I find myself enjoying these increasingly “liquid” boundaries in the Claude Code workflow, partly because my own interests happen to sit at the intersection of research, evaluation, data, quantitative metrics, and artistic creation.
This does not mean that everyone suddenly becomes an expert in every discipline. Rather, AI compresses many tasks that once required cross-team collaboration, long waiting cycles, or specialized execution into a space where one person can explore much more quickly.
As long as attention, time, and energy can keep up, one person can now move an idea from concept, to prototype, to a presentable demo with surprising speed.
This makes me think that one of the most important capabilities in the AI era may no longer be a single technical skill, but a new kind of work stamina: the ability to move rapidly across levels of abstraction, to translate vague ideas into executable workflows, and to switch between different modes of thinking without losing coherence.
Cross-functionality is no longer just about working with many teams.
Increasingly, it means becoming a small, multi-functional, constantly reconfiguring system yourself.