QFM107: Engineering Leadership Reading List - March 2026
Source: Photo by Vitaly Gariev on Unsplash
A short Engineering Leadership Reading List this month. The Mikado Method lays out how to make safe changes in complex codebases by visualizing dependency chains before cutting code, while Andrei Gridnev's Simple Framework for Software Engineering Management gives you a lightweight mental model for the people side. A practical walkthrough of 5 Claude Code skills shows how AI-assisted development fits into daily workflow, and ThoughtWorks gathers senior technologists at their Future of Software Development Retreat in Deer Valley to figure out what comes next.
As always, the Quantum Fax Machine Propeller Hat Key will guide your browsing. Enjoy!
Propeller Hat Key
- 0 of 5:
- Unrelated to technology, management, or leadership
- 1 of 5:
- Suitable for management and leadership novices
- 2 of 5:
- Topics of interest for the new manager or leader
- 3 of 5:
- Technology management and leadership in real-world use cases
- 4 of 5:
- Topics for experienced managers and leaders
- 5 of 5:
- Advanced topics in management and leadership
Links
The Mikado Method breaks complex refactors into small subtasks, tracked on a dependency graph. You make short timeboxed attempts (5-15 minutes), revert anything that fails, record what prerequisite you're missing, and try again. Only working code gets committed. The result is safe, incremental progress through legacy codebases without the usual all-or-nothing gamble.
Gridnev splits engineering management into three buckets: people (hiring, mentoring, reviews), delivery (planning, coordination, project tracking), and technical ownership. He tested the split with his own team for six months and found it clarified what he was actually supposed to be doing on any given day. Simple, but that's the point.
ThoughtWorks brought practitioners to Deer Valley to talk about what AI-native development actually means for engineering teams. The consensus: the hard work is shifting from writing code to supervising agents, reviewing their output, and catching what they miss. Security and org design haven't kept up. One provocative thread: source code itself may become a build artifact rather than the thing humans primarily work with, replaced by semantic layers that agents consume directly.
Regards,
M@
[ED: If you'd like to sign up for this content as an email, click here to join the mailing list.]
Originally published on quantumfaxmachine.com and cross-posted on Medium.
hello@matthewsinclair.com | matthewsinclair.com | bsky.app/@matthewsinclair.com | masto.ai/@matthewsinclair | medium.com/@matthewsinclair | xitter/@matthewsinclair
Was this useful?