Leadership in Tech
Leadership in Tech
#275 – January 08, 2026
from what you build to how teams work to why the work matters
Hello! 👋
This is the first Thursday issue. I hope you enjoy it and I'll manage to keep the pace up this year. There were loads of amazing articles over the Christmas break and the queue is strong.
The difference between manager and director
12 minutes by Nick Zylkowski
Nick shares his view how career progression from individual contributor to director follows a shift in focus from what you build to how teams work to why the work matters. He argues that early engineers focus on concrete outputs like code and features. Managers shift to improving team processes and effectiveness. Directors shape strategy and business outcomes, relying more on influence than technical expertise. Communication becomes the key multiplier for impact at senior levels.
Finally, a Real Workflow for AI-First Engineering
sponsored by Zencoder
AI can write code — but without structure, it creates more problems than it solves. Zenflow introduces a proper engineering workflow for AI: specs → plans → implementation → verification. Clean, enforceable, and built for multi-agent orchestration with parallel execution. This is how modern teams turn AI into a force multiplier.
Designing team traditions that fit your team
5 minutes by Christie Cosky
Team traditions don’t work the same way for every team. What feels fun and meaningful on one team may feel awkward or stressful on another. Christie explains why team culture matters when creating traditions and offers practical principles for designing ones that fit. When done well, team traditions can build trust, belonging, and connection—especially on remote teams—without needing a budget or a manager’s role.
AI systems engineering patterns
17 minutes by Alex Ewerlöf
Alex shares 30 AI Systems Engineering patterns from his 2.5 years of hands-on experience. He shows how traditional software engineering principles like caching and input validation still apply to AI systems. The patterns are organized into five layers covering interfaces, memory management, control flow, and advanced cognitive systems. The goal is to help senior engineers realize that AI development builds on existing engineering knowledge with some new twists.
Think of Pavlov
2 minutes by Andrew Bosworth
Most people approach conversations trying to win in the moment. But nearly every interaction is part of an ongoing relationship where you'll encounter the same people again. Each conversation trains others on how to work with you. If someone brings a question and feels dismissed, they'll stop asking. If you meet problems with curiosity, you'll get more challenging work. Your consistent tone and approach shape your reputation and long-term effectiveness.
Software acceleration and desynchronization
18 minutes by Fred Hebert
Software development involves interconnected loops of activities like coding, reviewing, and debugging. When teams use AI to speed up coding, they often find that code reviews become the new bottleneck. This happens because reviews serve multiple purposes beyond just catching errors. They spread knowledge, ensure quality standards, and connect different work cycles across teams. Simply speeding up one part can create mismatches with other parts that still need time to function properly.
Tests are dead. Meticulous AI is here.
sponsored by Meticulous
Meticulous automatically creates and maintains an exhaustive e2e UI test suite that covers every corner of your application – with no developer intervention required whatsoever. Dropbox, Lattice, Bilt Rewards and hundreds of organisations rely on Meticulous for their frontend testing. It is built from the Chromium level up with a deterministic scheduling engine – making it the only testing tool that eliminates flakes.
And the most popular article from the last issue was (unsurprisingly):