How to Perform Effective Project Management with AI
| Source: Towards Data Science
Tags: LLM agents, software engineering, productivity, prompt engineering, AI-assisted development
Software engineers are shifting from writing code (formerly ~70% of time) to prompting agents (~30%), testing (~30%), and exploration (~30%), according to a Towards Data Science tutorial outlining practical LLM-assisted project management techniques for the modern software engineer.
Details
As LLMs take over code generation, the role of the software engineer is transforming. Author Eivind Kjosbakken argues that the key productivity lever is no longer writing code — it's effective prompting and testing. He proposes that engineers now spend roughly 30% of their time prompting AI agents, 30% testing, 10% in meetings, and have 30% of capacity freed for other activities such as reading, exploration, and spinning up additional agents. The article covers practical techniques: how to structure tasks for agents, how to prioritize what to work on, and how to verify agent outputs through systematic testing. The framing is first-person and anecdotal — no hard productivity data is provided, and the time-distribution estimates are the author's own approximations, not empirical measurements. Most useful for engineers early in their AI-assisted workflow journey, offering a mental model for the new distribution of effort. The shift from execution to orchestration is the central insight: the bottleneck moves from writing to deciding, prompting, and validating.