Build software.
Understand the systems.
Own the outcome.
The work behind the code—engineering workflows, AI-assisted development, debugging, shipping, and building products under real-world constraints.
15 practical guidesOpen knowledge. No account. No fluff.
What are you trying to get done?
Start with the situation you have. Each route opens a practical guide, not a placeholder.
What are you trying to get better at?
Pick your situation. We'll sketch a practical route through the handbook.
I am preparing for my first engineering role
Learn the habits that help you contribute to a real codebase.
Your progress is saved on this device. No account needed.
Learn the work, not just the tools.
Software engineering is a chain of decisions. Start with the part of the chain you want to strengthen.
Engineering workflows
Turn vague tickets into safe, reviewable changes.
AI-native development
Use coding assistants without outsourcing your judgment.
Quality & debugging
Find the cause, not just a plausible-looking fix.
Delivery & operations
Ship safely and learn from what happens in production.
Startup engineering
Make good trade-offs when time and people are limited.
Engineering foundations
Learn the fundamentals that make every tool easier to understand.
Guides worth keeping open.
Written to be used mid-task: clear mental models, a repeatable workflow, and checklists you can take back to your editor.
Knowing the answer is good.
Knowing what to do next is better.
Real engineering is rarely a multiple-choice exam. But quick scenarios are a useful way to practice the instinct behind a good decision. Pick your move, get the debrief, and take the reasoning with you.
The agent says “all tests pass.” Now what?
A coding agent changed 14 files and reports that all tests passed. You did not see the test output. What is the most responsible next step?
TEST
A green check is useful only when you know what ran and what it covers. Inspect the diff and make the verification evidence legible.
Be the engineer who can own the outcome.
Start with one guide. Try one idea on a real task. Come back when the next question appears.