If AI writes code, what do senior engineers do

With the rise of AI and automation in software development, the traditional path to seniority and career growth for engineers is being compressed. Fresh engineers out of college can now tackle high-value business processes sooner, reducing the time spent on repetitive tasks. Kiran Kumar Katreddi, head of platform engineering and agentic AI at Meesho, explains that AI has significantly accelerated the rate at which engineers solve problems, leading to a shift in the maturity curve for seniority.

Senior engineers at Meesho are now more focused on systems-design thinking, as they design optimal architectures for thousands of microservices working together. While code writing is just a small portion of their work, they spend more time defining problems effectively. The scale of problems Meesho's engineers handle, such as processing billions of data points across multiple AI ranking models and running trillions of inferences daily, underscores the shift in focus.

AI and automation allow engineers to concentrate on scaling problems and unlock new possibilities, creating more interesting challenges for junior engineers as well. However, the claim that 70% of Meesho's code is AI-generated is somewhat ambiguous. Katreddi clarifies that they are referring to code generation, specifically the inner loop of the software development lifecycle, where developers build, test, and fix issues more efficiently with AI assistance.