Descriptions:
Recorded at The Pragmatic Summit, this panel conversation brings together Rajie (VP of Engineering at Atlassian) and Thomas (a newly announced startup founder) to examine what it concretely means to build an AI-native engineering team — not as an aspirational talking point, but as an operational reality they are both navigating.
Rajie describes Atlassian teams where engineers write zero lines of code manually, with work entirely handled through agent orchestration. He reports productivity gains ranging from 2x to 5x in typical cases and up to 100x in specific projects designed from scratch for AI-first workflows. A key structural observation is that the boundaries between PM, design, and engineering are blurring: PMs and designers are now writing and reviewing code, which is compressing feedback cycles and increasing creative output. The conversation also addresses where AI-native approaches break down — legacy codebases remain a hard constraint, with tools like Robo Dev struggling to maintain context and reliability on long-lived repos.
The panel reframes the common narrative around headcount reduction, arguing the more important question is what becomes buildable with AI that wasn’t feasible before, rather than how many engineers can be eliminated. Both speakers discuss the shift in engineering accountability from line-by-line code review toward defining guardrails, verifying inputs and outputs, and owning the correctness of agent-generated artifacts — a meaningful change in what senior engineering judgment looks like in practice.
📺 Source: The Pragmatic Engineer · Published February 24, 2026
🏷️ Format: Interview
