Michael Silvi

Principal Software Engineer

LinkedIn Profile

Michael has led several Stride engagements, including two organizational agile transformations and also a delivery project with 22 people across several teams. He kicked off an executive action team, a sales scrum team, coached a group of 20+ developers to improve communication and code quality, and been a core contributor to applications that have scaled to millions of concurrent users.

Posts by this author

Measure Fluency, Not Maturity

Maturity models rank you against a generic ladder. Why measuring code quality, delivery stability, and adoption together against your best teams is the real signal.

Individual AI Experimentation Doesn't Scale. Shared Rituals Do

Individual AI skill lifts individuals, not throughput. Why mob elaboration and mob construction turn one engineer's practice into how the whole team works.

Why Shorter AI Coding Sessions Produce Better Code

Model output degrades as context grows. Why bounding delivery scope and coding sessions with a plan-first discipline improves quality and cuts token cost.

Build a Knowledge Layer So Your AI Stops Relearning Your Business Every Session

Every AI coding session rebuilds context from scratch. A persistent knowledge layer of domain models and retrieval tooling encodes your intent so it doesn't.

Why Isn't Our AI Coding Tool Making Us Faster? Start by Structuring Your Repos

Your AI coding tool isn't the variable, your repos are. Why structuring repositories for AI readiness is the first move that makes every gain measurable.

The 2027 AI Budget Honesty Pass

Boards are asking what the AI budget actually bought. Run every line item through three questions — owner, metric, review workflow — before 2027 planning locks.