Gergely Orosz is the founder and author of The Pragmatic Engineer, an independent software-industry publication with more than one million readers, and the author of The Software Engineer’s Guidebook. A former Uber engineering manager, he investigates how technology companies build software, organize teams, measure performance, and adapt to AI-assisted development.
Orosz built trading software at J.P. Morgan, worked on Skype products at Skype and Microsoft, and joined Skyscanner before moving to Uber in Amsterdam. At Uber, he led mobile, web, and backend teams and worked on payments infrastructure. His engineering career gave him firsthand experience with distributed systems, hiring, promotions, and the organizational constraints shaping large engineering teams.
After pandemic-era layoffs disrupted his Uber team, Orosz left the company in October 2020. He had expected to finish a book and potentially start or join a platform-engineering company. Instead, he launched the paid edition of The Pragmatic Engineer in 2021, reporting on engineering organizations, compensation, infrastructure, management, and the software job market. The publication expanded into The Pulse, a research team, and a podcast, eventually reaching more than one million readers. It does not accept sponsorships.
Published in 2023, The Software Engineer’s Guidebook distills his approach to engineering careers: technical ability matters alongside communication, influence, production ownership, and understanding how organizations make decisions. Orosz self-published the book after more than four years of work.
- Productivity is not token consumption. Orosz challenges token-spending leaderboards and AI-usage targets that encourage engineers to generate visible activity instead of useful results. His reporting on AI developer-tool evaluation examines why mandates and adoption metrics cannot establish whether teams deliver better software.
- AI changes the whole engineering system. Faster code generation does not automatically remove review bottlenecks, unclear requirements, or coordination costs. His 2026 survey of software engineers, coauthored with Elin Nilsson, explores uneven benefits and rising costs; his AI Engineer Europe conversation also highlights the internal coding agents, service-discovery integrations, and developer platforms companies build around existing codebases.
- Product judgment becomes more valuable when code becomes cheaper. In his conversation with Linear cofounder Tuomas Artman, Orosz probes the consequences of accelerating feature production without understanding customer needs. Linear’s quality rituals and automated bug-fixing systems are Artman’s company initiatives; Orosz focuses on what they reveal about restraint, craft, and incentives.
- Agent orchestration is not people management. Coordinating coding agents resembles technical leadership more than managing employees: software does not require career coaching, conflict resolution, or the long feedback cycles involved in leading people. Orosz expects engineers to take broader responsibility for planning, product decisions, and operations.
His interview with database founder Simon Eskildsen applies the same scrutiny to infrastructure economics, caching, cloud capacity, and startup financing. The database architecture and company-building decisions belong to Eskildsen; Orosz draws out how technical simplicity, customer needs, and organizational choices determine whether ambitious engineering actually works.
▶ Watch ↗
