AI and engineering management: grounded in how teams work
Generative AI is valuable when it helps a team deliver better. My approach starts with engineering practices and keeps human judgment, quality and people development at the center.
Start with the real workflow
I begin with the steps where the team loses time or faces friction: code review, documentation and repetitive quality checks. The tool comes second; the team's needs and context guide how it is used.
Establish shared practices
At Witik, I introduced Claude Code to assist code reviews, structured prompts to make requests more consistent, and automated quality checklists. The goal is to make these practices useful to the team, not to multiply individual tools.
Keep responsibility with the team
AI can prepare and accelerate parts of the work; the team remains responsible for technical decisions and deliverables. The manager frames usage, supports skill development and protects quality throughout delivery.
Improve through concrete feedback
We look at the effect on review cycles and deliverable quality, then adjust our practices. In my experience, adopting these workflows reduced review cycles and improved deliverable quality.
